A PRACTICAL REVIEW OF THE EFFECTIVENESS OF 
ARTIFICIAL INTELLIGENCE IN THE AUTOMATED REVIEW OF 

LEGAL CONTRACTS. 

by 

Jacques Gerrit Fouché 

0516171M 

  

 

  

Submitted in partial fulfilment of the requirements for the degree of 

Master of Laws by Coursework and Research Report 

at the University of the Witwatersrand, Johannesburg 

 

 

Date: 17 October 2023 



 
 
 
 

DECLARATION 
 
 

I, Jacques Gerrit Fouché (0516171M), declare that this Research Report is my own unaided 

work.  It is submitted in partial fulfillment of the requirements for the degree of Master of 

Laws (by Coursework and Research Report) at the University of the Witwatersrand, 

Johannesburg.  It has not been submitted before for any degree or examination in this or any 

other university.   

 

I have submitted my final Research Report through TurnItIn and have attached the report to 

my submission. 

 
 
Word Count: 10,528 [Incl. the body of your report, footnotes BUT excl. title page, 
declaration, abstract, table of contents and bibliography]  
 
 
 
 
 
 
 
 
 
 
 
 



ABSTRACT

This research report compares the performance of AI-driven automated contract review 

platforms with that of human contract reviewers, aiming to see improved effectiveness for the 

elements of time, cost and quality. An empirical study is done by comparing the specific 

performance of an automated contract review platform provider, Lawgeex, to the human

contract reviewers of a business entity, Endress+Hauser. The results of the effectiveness

assessment are reported on through dashboard data and questionnaires to the users of the

platform. Recommendations are made both in general and specific to the two entities of the 

empirical study.



Table of Contents
I. INTRODUCTION ......................................................................................................... 2

II. LITERATURE STUDY................................................................................................. 3

(a) Fourth Industrial Revolution ......................................................................................... 3 

(b) Artificial Intelligence .................................................................................................... 4 

i. General Discussion ..................................................................................................... 4

ii. Current Use in Legal Profession ................................................................................ 5

iii. Potential Benefits of AI for Legal Profession .......................................................... 7 

iv. Potential Risks and Hinderances of AI for the Legal Profession ........................... 10

III. RESEARCH PROBLEM AND LIMITATIONS ........................................................ 11

(a) Artificial Intelligence in Automated Contract Review ............................................... 11 

(b) Hypothesis .................................................................................................................. 12 

(c) Limitations .................................................................................................................. 12 

(d) Artificial Intelligence Response to the Research Question ........................................ 13

IV. EMPIRICIAL STUDY ................................................................................................ 16

(a) General Description .................................................................................................... 16

(b) Data Assessment Parameters ...................................................................................... 16

V. DATA ANALYSIS REPORT ..................................................................................... 20

(a) Lawgeex Dashboard Data Analysis Report ................................................................ 20

(b) Target Group Data Analysis Report – Legal Counsel or Contract Reviewers ........... 27 

(c) Target Group Data Analysis Report – Contract Uploaders ........................................ 33 

(d) Data Analysis Discussion and Deductions ................................................................. 34 

(e) Data Analysis Conclusion ........................................................................................... 35

VI. RECOMMENDATIONS ............................................................................................. 35

(a) General ........................................................................................................................ 35

(b) Recommendations Specific to the Target Group ........................................................ 36

VII. CONCLUSION ............................................................................................................ 37

VIII. BIBLIOGRAPHY ........................................................................................................ 38

IX. ANNEXURES ............................................................................................................. 41

(a) Lawgeex Document Examples

(b) Questionnaire Formats and Analysis

(c) Lawgeex Dashboard Reports

(d) Respondents' Reports and Graphs

1



I INTRODUCTION

In the latter part of 2022, the public was introduced to the mainstream concept of artificial 

intelligence. The launch of the conversational application (app) called ChatGPT, put artificial

intelligence at the forefront of everyday discussions. However, artificial intelligence has been

in use for decades by data scientists and data engineers to process vast amounts of data. The

first use of the concept of computer intelligence was by Alan Turing when he published his 

work “Computer Machinery and Intelligence” in 1950.1 Computers can just process datasets

faster than humans can, with a lower margin of error. Tools such as ChatGPT and Google

Bard became mainstream since they could act conversationally on any topic, scourging the 

internet for answers. The user can very quickly gain insight into a topic that may previously 

have required significant time to research. In the business world, the use of artificial 

intelligence to process data is already long established. Consumer trends, financial market 

forecasting and similar large data sources have been analysed using computer processing

power. Human intelligence simply does not have the capacity or time to process the data

manually, with an acceptable low error percentage. Most often, the tasks are simply too

mundane to waste human resources on such an activity, when a computer can do it better and 

faster.

In the legal field, contract assessment often falls in the mundane category. Detailed contract

analysis in mergers and acquisitions is necessary to ensure that both parties are protected. 

Due diligence tasks are necessary to make accurate decisions. Many law firms today devote 

numerous hours of associates and paralegals, to review contracts and write assessments on 

which clauses should be revised or deleted, to provide a legal service to their clients. 

Artificial intelligence has entered this arena as well to provide faster service to legal 

practitioners. This research report will analyse the practical effectiveness (in time, cost and

quality) of an automated review of legal contracts by an artificial intelligence algorithm,

through the data and information of an artificial intelligence algorithm contract review 

service provider to a multinational business entity. In the business world, terms are often used

interchangeably such as “agreement” and “contract”. An agreement is considered a

negotiated and typically legally binding arrangement between parties as to a course of

1 Anon What is the history of artificial intelligence (AI)?

2



action.2 A contract is a written or spoken agreement, especially one concerning employment, 

sales, or tenancy, that is intended to be enforceable by law.3 Subsections of these would be a 

nondisclosure agreement and a supply agreement. A nondisclosure agreement (NDA) is a

contract by which one or more parties agree not to disclose confidential information that they

have shared as a necessary part of doing business together.4 A supply agreement is a contract 

between two parties that details the terms of a purchase of goods from a supplier.5 These 

subsections of agreements and contracts are what will be analysed in the empirical study.

This research report will juxtapose the performance of the automated contract review with

that of the human counterparts in terms of time spent, costs allocated and quality of the 

output.

II LITERATURE STUDY

(a) Fourth Industrial Revolution

The first industrial revolution was marked by the introduction of steam as a power source.

The advent of electricity, ushered in the second industrial revolution and the third was 

earmarked by automation and machinery. The Fourth Industrial Revolution was shaped when

computers breached the physical layer or “cyber-physical systems” made their appearance.6

The Fourth Industrial Revolution is characterised by many names or abbreviations – Industry 

4.0; Internet of Things; 4IR; IoT – but essential refers to the same thing.7 Smart digital

technologies are transforming the way companies do everything: design, innovate, improve,

manufacture and distribute their products to users and consumers, all over the world. These

technologies are disrupting the legal services industry as much as any other industry.8 

Industry 4.0 refers to a new phase in the Industrial Revolution where smart digital 

technologies such as automation, artificial intelligence, 3D printing, and the Internet of

Things are transforming the way companies manufacture, improve, and distribute their

products. The term was coined by Klaus Schwab, Founder and Executive Chairman of the 

World Economic Forum, in 2017 in his book The Fourth Industrial Revolution that a

2 https://languages.oup.com/google-dictionary-en/ accessed on 12.01.2023.
3 Ibid.
4 Ibid.
5 https://www.contractscounsel.com/t/us/supply-agreement accessed on 12.01.2023.
6 Anon What is Industry 4.0, the Fourth Industrial Revolution, and 4IR?
7 Ibid.
8 Anon Introduction to AI contract review Lawgeex.

3



revolution of a technical nature is underway "that is blurring the lines between the physical,

digital and biological spheres.”9 The Fourth Industrial Revolution builds on the digital 

revolution (The Third Industrial Revolution). Computers, electronics and the Internet from 

the 1950s to the 2000s are exponentially expanded by four basic types of disruptive 

technologies: 1) Connectivity, data and processing power; 2) computational analytics and 

intelligent computer systems (Artificial intelligence fits here); 3) advanced human-machine 

interfaces and 4) advanced engineering.10 These technological advancements are only part of 

The Fourth Industrial Revolution. Harnessing the power that these four disruptive

technologies bring is forcing businesses to rethink their operations and how to optimally use 

their greatest asset: their staff. Employees must learn new skills to use new tools that the 

Fourth Industrial Revolution brings to stay effective as a minimum or competitive and 

innovative as a norm. The legal profession is not exempt from this Fourth Industrial

Revolution.11 Legal practices today deal with many contracts in various forms. Mergers and

acquisition departments, due diligence assessments and the like are all grounded in contracts. 

These contracts of clients need to be scrutinised by the legal practice to provide professional 

advice to clients. The contracts should be analysed by experienced legal professionals such as 

associates and paralegals to compile comprehensive feedback reports that can be used by the 

partners in law firms to brief clients. This contract review is a time-consuming task but needs

to be done with accuracy, despite the nuances of contract language. Mistakes will inevitably

happen due to the nature of the task – being mundane and repetitive – but there are tools 

available that can assist the legal practice. The Fourth Industrial Revolution bring

computational analytics and advanced human-machine interface that are aimed at alleviating

the mundane task and providing accurate results.

(a) Artificial Intelligence

i. General Discussion

Artificial Intelligence (AI) explained in broad terms would be when an algorithm, based on a 

computer system, completes a set of tasks normally done by humans. This includes analysis, 

strategies and conclusions on a set of facts and compared to a defined set of rules. AI is new

in the greater scheme of things but such ability has been theorised about for many years.12 A

9 Schwab, Klaus The Fourth Industrial Revolution.
10 Op cit note 7.
11 Ibid.
12 Vanni, Rachel How Artificial Intelligence Is Transforming the Legal Profession.

4



German lawyer, G.W. Leibniz, put a theory forward in the late 1600s that a futuristic

machine would be able to calculate an output, based on a set of rules and a defined input. He 

never saw any “machine” that remotely resembled a computer, but he was accurate in the

benefits that such a process would provide to humans and indeed lawyers. He prophesied that

“excellent men would relegate calculation tasks to machines” to focus on more important 

tasks.13 Futurists and experts alike see the impact of AI as important as the innovation of 

steam.14 The impact will reverberate throughout all industries and professions, yet there is a 

fair amount of caution about the abilities of AI. Renowned physicist, Stephen Hawking, had

candid responses to an online question-and-answer session called “AMA” or “Ask Me 

Anything”. He cautioned that super-intelligent AI would be beneficial provided that its goals 

were aligned with ours.15 Further, that human-level AI is a matter of time and we need to 

ensure that the AI is creating beneficial intelligence for us.16 Lastly, when AI will have higher 

intelligence than its creators, we will face an intelligence explosion bigger than what we can 

currently comprehend. Machine’s intelligence will surpass ours more than our intelligence

surpasses that of snails.17 This is called “the singularity” where the advances in technology

will cause machines to be smarter than human beings.18 But, let me leave the super-futuristic 

concepts for another time.

ii. Current Use in Legal Profession

In the legal profession, AI and humans can and have collaborated to improve the 

effectiveness – time, cost and quality – of legal functions. AI can analyse and summarise 

pages upon pages of legal precedents, can review and mark up contracts and can search

through mountains of legal data to find answers. All being done in a matter of hours.19 For

reference, LexisNexis alone processes 13 million new documents daily from over 50,000 data 

sources in varying formats. Its legal big data platform holds 60 billion documents and 2.5 

petabytes of legal data.20 The ‘big data’ referred to is extremely large data sets that may be 

analysed computationally to reveal patterns, trends, and associations, especially relating to

human behaviour and interactions.21  Although humans can work through such data volume,

13 Wolfram, Stephen Dropping In on Gottfried Leibnitz.
14 Anon Lawyer vs AI: A legal revolution.
15 Tegmark, Max Hawking Reddit AMA on AI.
16 Ibid.
17 Ibid.
18 Reedy, Christianna Kurzweil Claims That the Singularity Will Happen by 2045.
19 Op cit note 15.
20 Pfeifer, Jeff LexisNexis: Legal AI and Our Data-Driven Future.
21 Op cit note 3.

5



the time taken to do so may render the answer obsolete by the time it is found. The 

capabilities of AI must wow even the uninformed and casual observer but a topic that can

never be overlooked is ethics. Once a system is acting intelligently and displays distinct

autonomy, crucial legal and ethical questions are raised where the answers impact both the 

creator and consumer of AI technology. Such questions include public policy, professional 

ethics and philosophical ethics and the answers need to come from a combination of

computer scientists and legal professionals.22 For example, if automated vehicles reduce

traffic fatalities by half, would the car makers receive thank-you notes or lawsuits? Would it 

be ethically acceptable that a machine decides a lesser probability of human injury versus a 

greater probability of material damage?23 A real threat is that AI can harm by perpetuating 

human biases.24 Think of ethical profiling through facial recognition technology. Think 

invasion of privacy through digital video technology under the auspices of targeted 

advertising. AI, the law and ethics are a crucial short-term research priority defined by Stuart 

Russell, Daniel Dewey and Max Tegmark and will uncover more areas as AI develops.25 An 

important subsection of AI and its ability to act intelligently is machine learning. Due to the 

advances in computer hardware and software, machine learning has expanded rapidly. 

Machine learning is where computer algorithms learn independently from the data and 

information available. Computers autonomously change and improve their algorithms 

themselves. It is already used today for self-driving cars, creating sports reports or locating 

persons of interest. The advancement in computational power allows machine learning to 

improve faster and provide results with greater accuracy. These are natural progressions in

any industry based on a vast body of knowledge. For data to become information, it must be

standardised, classified, summarised and stored in an accessible format. This is ideally suited 

for AI and machine learning. These features are already used in the legal industry to remove

the mundane tasks from expertly trained lawyers, allowing them to practice law.26 The First

Industrial Revolution automated manual labour. In the same way, the Fourth Industrial 

Revolution is automating mental tasks.27 This has a definitive effect on the practice of law.

AI will aid legal practitioners rather than replace them in the foreseeable future. Contract

review, document discovery and legal research are all part of the current AI capabilities. 

However, AI has already started with contract drafting, legal outcome predictions and

22 Stuart, Dewey, and Tegmark Research Priorities for Robust and Beneficial Artificial Intelligence.
23 Ibid.
24 Anon Artificial Intelligence Law.
25 Op cit note 23.
26 Anon Big-Firm Intelligence for Small Law Firms page 4.
27 Stepka, Matthew Law Bots: How AI Is Reshaping the Legal Profession.

6



recommendations on judicial decisions.28 AI offers benefits to the legal practitioner – both

current benefits and potential benefits. Increases in legal practitioner productivity and

improved accuracy on repeated legal clauses are already a reality. Yet AI is not yet 

sufficiently developed to replace human judgement in the legal sphere. The high risk of bias

used in setting the base algorithm remains real in the development of AI as well as the

“explainability” of AI decisions to humans. These must be overcome first before a deeper 

entrenchment in advanced legal contexts can happen.29 Some legal practitioners and business 

entities are already using AI and machine learning in contract review for a quicker, more 

accurate outcome. The ability of AI to spot incompatible clauses and errors, more 

consistently than its human counterparts, is already providing entities with improved 

efficiency. Start-up businesses like Lawgeex are already providing such services in a

sustainable business model.30 Traditional lawyers are rightfully cautious about the

encroachment on a traditionally human area of expertise. But the objective of AI is not to 

replace legal practitioners. It aims to automate repetitive tasks that require time and high

accuracy from their human counterparts but is not necessarily very exciting. This allows the

same legal practitioners to focus on detailed assessments and personal contact with clients.31 

AI and machine learning will without a doubt create new career paths and direction in the 

legal sphere.32 AI’s strength lies in the superior ability to process data but human traits of 

emotional intelligence and human judgement is (currently) beyond the reach of AI. The high

complexity found in tax law will gain from AI’s processing ability, but humans are superior

in people interaction areas – negotiation, mediation and ethical behaviour.33 Just as the

plough didn’t make the farmer redundant; it improved their effectiveness. AI is an instrument 

that makes the lives of users easier by removing mundane tasks and allowing them to focus 

on their target market – people.34

iii. Potential Benefits of AI for Legal Profession

The effectiveness of AI is evaluated for time, cost and quality. A legal practitioner’s ability to 

earn is directly linked to billable hours. AI drives the tools that assist legal practitioners to 

make better legal decisions, faster. Such decisions can be encoded as an algorithm for future

28 Op cit note 28.
29 Ibid.
30 Ibid.
31 Op cit note 13.
32 Ibid.
33 Anon Lawyer vs Artificial Intelligence: A Legal Revolution.
34 Op cit note 9.

7



machine learning, meaning that the corporate legal practitioner does not need to be 

everywhere at the same time. His knowledge is duplicated for use within his organisation,

supported by AI.35 The mobile world required people to be available whenever their clients

need them. Videoconferencing from any location; taking calls while on the daily commute; 

etc. are all ways in which our digital world has improved our accessibility.36 It is estimated

that 23% of an average legal practitioner’s day can be freed up with AI and automation of

tasks. The “paperwork-based” duties are taking away quality time from more important 

aspects of a legal practitioner.37 The review of documents in due diligence or litigation, 

contract clause assessment and qualifications are already done consistently with AI.38 As the

legal industry grows and develops with the aid of AI, the know-how of practitioners is more

important than standard legal practice. There are already trends that corporate entities are no 

longer willing to pay for the time spent by associates on contract review, because of the 

abilities of AI in the same field. Such shifts are opening the door for law firms to increase 

fees for experience engagements with clients such as consultations and assessments.39 

Corporate entities dealing with high-value or high-risk opportunities need to have legal 

answers quickly before opportunities disappear or risks materialise. The AI-assisted legal 

practitioner thus acts with sensitivity towards the business.40 Increased data processing speed 

leads to a change in cost. Previously, such hardware was located at a client’s site, and this had

to be maintained by IT specialists. Improvements in technology now leave the hardware at

the service provider’s site and data is uploaded to the cloud via the internet. Legal 

practitioners are seeing cost reductions in client representation and practice management. 

Nearly all documentary processes are becoming digitalised and automated at a fraction of the

cost of the same activity by people.41 Start-up enterprises such as Lawgeex have a yearly

license fee of about $75,000, including the roll-out of the assessment criteria (“platform 

playbook”) within the service and the training of the users on the system.42 It was already

stated that AI can analyse and summarise huge amounts of legal data without the help of a

legion of associates.43 Legal practitioners can spend time on the complex side of their 

business – argument formulation, negotiation, client interaction and strategic planning –

35 Op cit note 9.
36 Op cit note 7 page 7.
37 Anon How Will Artificial Intelligence Impact the Legal Profession?
38 Ibid.
39 Ibid.
40 Op cit note 9.
41 Op cit note 34.
42 Fischer & Cohen The Total Economic Impact™ of Lawgeex page 2.
43 Op cit note 34.

8



knowing that AI will deliver.44 The baseline work requirement of all law firms is to assess 

agreements, whether they are non-disclosure agreements (NDA), supply agreements (SA) or 

service level agreements (SLA). These contracts are engaged multiple times per year with 

multiple clients, yet the contract clauses are not always unique or agreement specific. The

standard clauses have come past in one iteration or another. Rather than reviewing these same

clauses of the same agreements – day in and day out – by a human legal practitioner, AI can 

review these agreements without tiring and without errors. Should the same mistake be 

repeated, it will identify the mistake every time without wavering accuracy. Should 

contracting parties employ ambiguous wording, AI will highlight and identify these for 

changes to clearer wording.45 The accuracy of the work delivered by an AI contract review

platform creates unprecedented contract quality. In corporate entities, the business units 

relying on the legal department for agreement review can be sure that all issues have been 

identified, addressed and amended to company standard wording. This gives confidence to 

such business units to continue with their business activities. Coupled with the turnaround

speed of typically less than 48 hours, both time and quality goals have been achieved.46

Ultimately, the goal of AI development is to reach the ability of systems to learn like humans 

and surpass their ability in most analytical tasks, thus positively influencing society. The 

quality delivered by AI will consistently improve with the ability of AI to learn.47 

Collectively, the effectiveness of AI, measured in time, cost and quality, leads to gains in the

efficiency of the legal practice, corporate entity, or business unit. This application of AI helps

small law firms to grow, increasing their efficiency to focus on clients, sales and operations.48

The efficiency of AI lies in its ability to be programmed to copy human analysis and decision

criteria. It can recognise patterns from as little as 20 legal documents or 20 million

documents. The efficiency of the AI system is its speed to identify patterns. Whether the 

source documents are judicial reports or contract documents, the AI system can analyse, 

summarise, draw conclusions and provide a human-understandable output format.49 Machine 

learning further enhances its ability to analyse and data mine through repetitive use, further 

aiding its human counterparts.50  This technology is available to any legal practice, whether a

44 Op cit note 13.
45 Op cit note 9.
46 Ibid.
47 Op cit note 28 page 109.
48 Op cit note 27 page 3.
49 Ibid page 11.
50 Op cit note 27 page 11.

9



single attorney for a large corporate.51 AI reduces the historical advantage of large law firms 

with legions of hired associates since AI can be used by anyone.52

A recent LexisNexis report on the evolution of legal research identified six ways that AI is

transforming legal practice. Those six categories are 1) Due Diligence; 2) legal analytics or 

prediction technology; 3) document automation; 4) intellectual property, 5) electronic billing 

and 6) visualisation tools.53 Note that document automation – the use of AI to create or 

complete documents – is not automated contract review. This is creating or completing a 

document by filling in predefined fields that fall outside the scope of this report. Automated

contract review forms part of the due diligence category and this report will only focus on

automated contract review with AI. Agreements and contracts received by firms remain an 

obstacle to business and legal entities. Notwithstanding the volume of these, the myriad of

versions, use of language and challenges to control these contract documents. AI together

with machine learning can be programmed with parameters to search, identify, check and

update similar concept clauses automatically.54 Firms can now manage contracts through the 

algorithm that identifies key components and supplies basic predictive alternatives. The AI

can now, based on the company-specific contract clause criteria, assess any contract and

provide alternative clauses in the case of conflicts, flag unacceptable clauses and make 

recommendations in case of ambiguous language.55 The negotiating legal practitioner is 

immediately armed with clauses to accept, amend or reject – far sooner than with a 

conventional contract review. As the AI body of knowledge grows, the AI software can draw 

conclusions, develop contract strategies and discover novel interpretations.56

iv. Potential Risks and Hindrances of AI for the Legal Profession

At an almost basic level is the answer to the question “What is AI?”. The process or software 

or algorithm of 10 years ago is not the same as today. It can best be described as a constantly 

moving target. The question of “What constitutes intelligence?” is equally hard to define.57 

The law functions on a set of rules that demarcate an action or a concept. There is a legal

definition for terms such as “contract” or “breach”, yet the definition of AI is opaque at best.

There are several risks to the use of AI in the legal profession. As it is a software program,

51 Ibid page 7.
52 Ibid.
53 Anon Six Ways Artificial Intelligence (AI) Is Transforming Law Practice.
54 Op cit note 34.
55 Op cit note 54.
56 Op cit note 13.
57 Greenstein, Stanley Preserving the Rule of Law in the Era of Artificial Intelligence (AI).

10



arguably not in the traditional sense, its functioning can be impeded by coding failures. 

Further, as the AI training is based on basic data or training data, the algorithms could be

overfit or underfit, thus skewing results. The results can even be irrational. The AI algorithm

may potentially be unusable on certain data sets where the training data were inadequate. 

Lastly, AI systems could actively discriminate against people, ideologies or orientations 

whilst carefully duplicating and advancing the current bias of society.58 Practical examples 

would be a persistent rejection of any African country’s law as the governing law of a 

contract or the continuous rejection of established seats of arbitration in place of foreign seats 

of arbitration. There are several examples of where AI systems made “wrong” decisions. A 

self-driving vehicle crash in 2016 claimed the life of a pedestrian and the Microsoft chatbot 

“Tay” responded with racist and sexist responses and even ended with denying that the

Holocaust happened. The chatbot was shut down after a mere 16 hours of operation.59 Who 

would be held to account? There is no natural person or juristic person that can be sued since

the response originated from an entity separate from its creators. Hence the responses of such

systems should be carefully scrutinised to ensure correctness.

III RESEARCH PROBLEM AND LIMITATIONS

(a) Artificial Intelligence in Automated Contract Review

The literature study identified the impact of AI on the legal profession. Its ability to rapidly 

analyse data and trends is established. Artificial intelligence tools and software are used as a 

blanket term for the function of automated contract review. However, the aim of this report is

not to define AI or deep dive into the inner workings of the AI service provider’s intellectual

property (IP). Suffice it to say that for this research report, “a computer programme” or 

“software” compares an uploaded contract against a set of predefined, acceptable contract

clauses and identifies those clauses which are not compliant, too ambiguous, or unclear. The

set of predefined, acceptable contract clauses is called a “playbook” by the AI platform

service provider. The AI will render an output document that either “strikethrough” the 

incompatible clauses without a reworded clause or will add additional descriptive clauses, to 

define the limits of risk and liability. In the case of general clauses, the AI will add comments 

to the document to aid the receiver in further contract negotiation. This research report will

58 Kaminski, Margot E Regulating the Risks of AI.
59 Buiten, Miriam C Towards Intelligent Regulation of Artificial Intelligence.

11



analyse if the expected benefits, measured as time cost and quality, are realised in practice. 

Data sets from an AI service provider and their customer will be used to determine if the

benefits are realised. First, the evidence of AI technology applied to contract review will

determine if any efficiency gains are realised. The efficiency will be analysed as a function of

time. Secondly, the costs associated with the AI application used for contract review will be

countered by human contract reviewers. A licence cost versus an hourly cost for the same

functions will be used to determine a cost-benefit. Thirdly, the actual revised clauses of the 

contract updates as done by the AI application, will be measured against the human 

reviewers’ output and a quality review will thus be done. This measures the accuracy of the

AI application review when compared to human reviewers. It is essential to keep in mind that

the empirical study is limited to the automated contract review of an artificial intelligence 

application.

(b) Hypothesis

The overarching hypothesis is that AI contract review is more effective in terms of time, cost 

and quality than traditional manual contract review. The three segments will be analysed

individually to arrive at the answer to the overarching hypothesis. Sub-hypothesis 1: AI

contract review is more efficient, i.e., faster than human reviewers. Sub-hypothesis 2: AI 

contract review is more cost-effective, i.e., cheaper than human reviewers. Sub-hypothesis 3: 

AI contract review yields higher quality, i.e., fewer mistakes and more consistent responses, 

than human reviewers.

This research report will compare the two methods (AI contract review and human contract 

review) based on real-life data and through actual reviewers’ feedback to either prove or 

disprove the hypothesis.

(c) Limitations

The real limitation is that the performance of the AI platform is limited to the information 

entered by the employees of the Endress+Hauser Group. No other data from within the

Lawgeex platform was brought into the mix. Thus, machine learning (if any) is limited to the

data of the Endress+Hauser Group. The data within the AI platform has a maximum span of a

rolling 12-month period. Therefore, the contracts within the system are only those uploaded 

in the data frame. Specific details about this will be given in the data analysis report.

12



(d) Artificial Intelligence Response to the Research Question

This report would be lacking input if modern conversational apps such as ChatGPT are not 

polled for a response. I opted to pose the research question to three different conversational 

AI apps: ChatGPT, Google Bard and Shako. The responses are unedited.

Question: Is legal contract review by artificial intelligence more effective in terms of time, 

cost and quality than traditional manual contract review?

i. Chat.openai.com (ChatGPT)60

“Legal contract review by artificial intelligence (AI) can offer several advantages over 

traditional manual contract review in terms of time, cost, and quality. Here are some ways in 

which AI can be more effective:

Time Efficiency: AI can process and analyze contracts much faster than humans. It can 

review large volumes of contracts in a fraction of the time it would take a human to do the 

same task. This speed allows for quicker turnaround times and enables legal professionals to 

focus on more complex or strategic aspects of their work.

Cost Savings: AI-powered contract review can significantly reduce costs compared to manual

review. By automating the review process, organizations can save on labor expenses and 

allocate resources more efficiently. Additionally, AI can help identify potential risks, 

inconsistencies, or missing clauses, reducing the likelihood of costly errors or legal disputes 

down the line.

Quality and Accuracy: AI can enhance the quality and accuracy of contract review. Machine 

learning algorithms can be trained on vast amounts of data, including legal precedents and 

best practices, enabling them to identify important clauses, potential issues, and legal risks. AI 

can also help ensure compliance with regulatory requirements by flagging any deviations or

non-compliant clauses.

Consistency and Standardization: AI-based contract review ensures consistent application of 

legal rules and standards. It eliminates the variability that may arise from human 

interpretation and subjectivity, providing a standardized approach to contract analysis and 

reducing the chances of inconsistencies across different reviewers or departments.

60 ChatGPT, available at https://chat.openai.com/, accessed on 18 July 2023.

13



Scalability: AI can easily scale to handle a large number of contracts simultaneously. Whether

you have dozens or thousands of contracts, AI can process them all efficiently and 

consistently, without compromising quality or incurring additional costs.

Continuous Learning: AI can continuously learn and improve over time. As more contracts 

are reviewed and new legal precedents emerge, AI algorithms can adapt and incorporate this 

new knowledge into their analysis. This adaptive learning enables AI to enhance its 

performance and accuracy, making it increasingly effective over time.

It is important to note that AI should be seen as a valuable tool to assist legal professionals, 

rather than a complete replacement for human expertise. While AI can automate routine tasks 

and streamline the contract review process, human oversight and judgment are still essential,
particularly for complex legal issues or unique contractual arrangements.”

ii. Bard.google.com61

“Yes, AI-powered legal contract review is more effective in terms of time, cost, and quality 

than traditional manual contract review.

• Time: AI can review contracts much faster than humans. This is because AI can

process large amounts of data and identify key terms and clauses much more quickly 

than a human reviewer. For example, a study by Gartner found that AI-powered 

contract review can reduce the time required to review a contract by up to 50%.

• Cost: AI can also help to reduce the cost of contract review. This is because AI can

automate many of the tasks that are currently performed manually, such as data 

extraction and keyword searching. For example, a study by CobbleStone Software 

found that AI-powered contract review can save businesses up to 30% on the cost of 

contract review.

• Quality: AI can also help to improve the quality of contract review. This is because

AI can be trained to identify potential risks and errors in contracts. For example, a 

study by Lexology found that AI-powered contract review can help to identify 

potential risks in contracts up to 90% of the time.

In addition to the benefits of time, cost, and quality, AI-powered legal contract review can 

also help to improve:

61 Google Bard, available at https://bard.google.com/, accessed on 18 July 2023.

14



• Compliance: AI can be used to help businesses ensure that their contracts are

compliant with all applicable laws and regulations.

• Negotiation: AI can be used to help businesses negotiate better terms in their

contracts.

• Risk management: AI can be used to help businesses identify and mitigate potential

risks in their contracts.

Overall, AI-powered legal contract review is a more effective way to review contracts than 

traditional manual contract review. It is faster, cheaper, and more accurate, and it can also 

help businesses to improve compliance, negotiation, and risk management.

However, it is important to note that AI-powered legal contract review is not a perfect

solution. AI systems can still make mistakes, and they can only be as good as the data they 

are trained on. As a result, it is important to use AI-powered legal contract review in
conjunction with human review.”

iii. Shako.ai62

“Hello! As Shako, I can tell you that there are benefits to using artificial intelligence (AI) for 

contract review. One of the main advantages is that AI can process large amounts of data 

much faster than humans, which can save time and reduce costs. Additionally, AI can identify

patterns and anomalies in data that may be difficult for humans to detect, which can improve

the accuracy of contract review.

However, it's important to note that AI is not a replacement for human expertise in contract

review. While AI can be helpful in identifying certain types of issues, it may not be able to 

catch every mistake or fully understand the nuances of a contract. Additionally, there may be 

cases where a human review is necessary for legal or ethical reasons.

Overall, I would say that AI can be a useful tool in contract review, but it should be used in 
conjunction with human expertise and not as a replacement for it.”

62 Shako, available at https://shako.ai/, accessed on 18 July 2023.

15



IV EMPIRICAL STUDY

(a) General Description

I am a long-term employee of a multinational company and have worked in a variety of roles 

within the organisation, both locally and abroad. A career that started as a chemical engineer, 

developed into a sales and marketing function, and later specialising in project management. 

This brought me into active contact with contracts and organisational development initiated 

the use of automated contract review. During this time, I was prompted to pursue a law 

degree to become a more skilful contract negotiator. The love of studies and self-

improvement led me to a Master of Laws degree with a specialisation in commercial and

business law. My current role includes the function of corporate legal counsel and gives me 

insight to the requirements of contract review in a multinational organisation.

Endress+Hauser is a multinational business entity that receives a variety of contracts and 

agreements during its business dealings. Among these are non-disclosure agreements and

supply agreements (contracts). I will expand on their business functions when describing the

platform users.

Lawgeex is a service provider of automated contract reviews by using artificial intelligence 

software, based on a defined set of rules provided by a business entity. Endress+Hauser is 

such a business entity and Lawgeex is such a service provider to them. Lawgeex self-identify 

what they do as:

“We use Artificial Intelligence to relieve corporate attorneys from the daily tedium of 

contracting; with more time and creative energy, they can use their expertise to add more 

value and make a bigger strategic impact on their business. We’ve reviewed millions of 

contracts, expanding our AI’s capabilities to match the output of a human lawyer. Lawgeex 

harnesses contextual intelligence to understand legalese, accurately reviewing contracts, 

improving negotiation and accelerating approval at scale. The document is uploaded to the 

platform and the review parameters are set for location and document type.”63

The focus of this research report is to determine the effectiveness of AI in automated contract

review, by analysing the aspects of time, cost and quality and to juxtapose that with the 

performance of the same aspects of human contract reviewers. The AI platform has provided

63 Anon “What we do” available at https://www.lawgeex.com/company/, accessed on 11.01.2023.

16



me access to its dashboard for all aspects. I have created a questionnaire for the human 

participants, analysing their responses according to selectable pre-defined options.

The participants come from the entire Endress+Hauser Group and I assumed that contract 

review is done in the same manner across the Group. The various entities have different legal

systems that may impact their natural decisions but the product offered and the customer type

worldwide is similar. I have also assumed that the aspects are known and visible to the

participants in that they can accurately respond to the time and cost questions. The quality 

aspect will be a comparative study between the AI platform dashboard and that of the human

respondents and I expect to see a close similarity between the AI performance and the human

reviewers.

A baseline was determined by asking the contract reviewers within the target group how

contract review was done before the introduction of the AI service provider’s platform. A 

questionnaire64 with selectable pre-defined answers, relating to effectiveness, was provided

and the responses were analysed.

(b) Data Assessment Parameters

i. Target Group

The target group for the empirical study centred on two subgroups, namely those who 

uploaded agreements (they were called “Uploaders” in the questionnaires.) and those who 

conducted reviews of the replies received (These were called “General Counsel, Legal

64 Annexure: Endress+Hauser Legal Counsel Questionnaire.

17



Counsel, Legal Support and Paralegals” or “Legal Counsels” in the questionnaires.). The

“Uploaders” questionnaires were sent to everyone who uploaded an agreement to Lawgeex in

2022. This was a total of 74 people across 23 countries. The “Legal Counsels” questionnaire

was sent to all staff with such designation in their job title. This was a total of 32 people 

across 9 countries.

ii. Explanation of AI “platform”

Lawgeex possesses a patent for “System and Method In Support of Digital Document 

Analysis” and its proprietary algorithm that analyses contracts in a contextual format. It is 

done by a learned understanding of contract language as well as the policies and objectives of 

the parties in a contract. Lawgeex can uniquely locate and respond to modulations, negating a

compromise on quality assessment speed and costs of AI contract review.65 Lawgeex

achieves this with its “Contextual AI Engine” that understands contract language as well as 

the intent and objectives of the parties in a contract. It aims to mimic the approach used by

human lawyers to always consider the client’s guidelines and goals. As the contract body of

knowledge grows, so new patterns are identified and machine learning is employed to grow

the contextual know-how.66 Lawgeex is considered a “pioneer” in this field claiming a 10x 

faster speed than a human lawyer, with a better cost structure and higher quality.67

Lawgeex works on a legal “playbook” that holds the company’s legal policies and preferred

clause language. It also contains the specific risk assessment and risk appetite of the 

company. Those legal policies must be applied consistently for them to be truly protected. 

When the AI engine is programmed with the company’s legal playbook, it can analyse any 

3rd party contract, without prior qualification.68

When a user uploads an agreement to the Lawgeex portal, both the location country and the 

playbook need to be selected by the user. This determines both the applicable location law as 

well as the set of rules to be applied to the uploaded agreement. The user-friendliness, any 

pdf, doc or docx format can be uploaded to the portal. The AI-assessed result is returned to 

the user in MS Word format with all deletions and additions marked red and, in some 

instances, additional side comments are added. See ANNEXURES for examples of a base 

contract and a marked-up contract.

65 Anon Lawgeex granted pioneering patent.
66 Ibid.
67 Ibid.
68 Op cit note 9.

18



A dashboard is made available to the senior personnel of the corporate entity to monitor and

track specific analytics regarding the performance of the Lawgeex portal. I received special 

permission to access this information for this research report.

iii. Explanation of the “platform users”

Endress+Hauser self-identify as:

“…a global leader in measurement instrumentation, services and solutions for 

industrial process engineering. They provide process solutions for flow, level, 

pressure, analytics, temperature, recording and digital communications, optimizing 

processes in terms of economic efficiency, safety & environmental impact. Their 

customers come from various industries, including chemical, food & beverage, life 

sciences, power & energy, mining, minerals & metals, oil & gas and water & 

wastewater.”69 The Group has a variety of offices around the world, functioning as 

either a sales organisation or a production organisation.”

Endress+Hauser is a corporate client, utilising the Lawgeex platform for the automated 

contract review of non-disclosure agreements and supply agreements. The Endress+Hauser

Group General Counsel developed a set of minimum rules that is acceptable to the Group to

harmonise the contractual clauses throughout the Group. Cognisance is given to common law 

and civil law principles as well as to international contract law.

Within the Endress+Hauser sales organisation, there are several departments – General 

Management, Sales, Projects, Finance or Controlling – fulfilling specialised roles in the 

greater functioning of the company. The Sales department deals with customers on a 

transactional basis for daily maintenance and repair orders (MRO). The Projects department

deals with customer projects where engineers, project managers and digital system engineers

collaborate to support a customer project to build a new facility – mine, brewery, refinery –

etc. The latter requires a larger number of items to be delivered at a specific time, hence the

monetary spend is higher. Because of this, legal agreements such as non-disclosure 

agreements and supply contract agreements become necessary.

Contract review within the company starts when an employee receives an agreement from a 

customer. The document is uploaded to the platform and once a response is received, it is

69 Anon About us: The Endress+Hauser Group.

19



reviewed by either legally trained staff or senior management. The final reviewed document 

is then returned to the customer as a baseline document for negotiation.

iv. Empirical Research Methodology

The research methodology employed in the empirical study was to engage with two groups of

platform users – the legally trained staff and the general staff tasked with uploading 

agreements – to complete questionnaires regarding their work. The questions overlapped 

extensively to determine the understanding and accuracy of the replies received from the 

platform.

The platform is also capable to provide reports regarding several parameters. I received 

special access to the back-end reports to tailor my questions to the platform users and to gain 

operational insight into the AI working. The data made available to me was limited to the 

data entered by Endress+Hauser into the platform but from any user, anywhere in the world 

where the company operates.

The data of the questionnaires were compared to the analytics on the Lawgeex dashboard to 

find similarities and discrepancies. Deductions were made from these to find the reasons and 

develop recommendations. Additionally, specific company internal questions were asked of 

all respondents to provide specific data regarding costs. This is not captured by the Lawgeex 

platform but was deemed relevant for the research.

V DATA ANALYSIS REPORT

(a) Lawgeex Dashboard Data Analysis Report

The Endress+Hauser Group provides access for any employee to the Lawgeex platform for

the upload of non-disclosure agreements and supply agreements. In this report, I will refer to

them as “Uploaders”. The data period under assessment (9 January 2022 to 9 January 2023)

saw 89 different people uploading agreements from 24 different countries. A slight 

complexing factor was that several countries with Endress+Hauser offices have a business 

responsibility for a different area, region or country. For example, Endress+Hauser South 

Africa has business responsibility for sub-equatorial Africa with a northern territory border of 

Kenya, Uganda and the Democratic Republic of the Congo. All the countries south of this 

boundary line have existing and potential customers of Endress+Hauser South Africa. Should 

contracts be uploaded by a South African employee, the uploader can select the country of

20



origin of the contract. In the dashboard data, however, this location selection is not 

transparent.

The Lawgeex dashboard provides four reports in both graphical pdf format and base data MS

Excel format. Report 1 covers the Legal Operations, with a high-level view of the

performance of the Endress+Hauser, uploads to the Lawgeex platform. Report 2 covers the 

Legal Policy: Mutual non-disclosure agreements, Report 3 covers the Legal Policy: 

Receiving non-disclosure agreements and Report 4 covers the Legal Policy: Supply

Agreements. I have excerpted some of the graphs for illumination in the text but the complete

reports are added to the ANNEXURES. For clarity, the difference between the two non-

disclosure agreements is that one (mutual) relates to both parties disclosing information and 

the second (receiving) relates to only Endress+Hauser receiving information from an entity. 

There are also different playbooks for Supply Agreements such as “General” or “Common 

Law”, but the evaluation does not need to delve into such differentiation. In the sections 

below, the key graphs are duplicated to provide clarity within the text.

i. Lawgeex Report 1: Legal Operations70

This covers the period from 9 January 2022 to 9 January 2023 (note 1) with a total of 443 

contracts uploaded (note 2). The average turnaround time was 1 day, 1 hour and 49 minutes

per contract (note 3). There seems to be a seasonal trend where fewer contracts are uploaded

during the Northern Hemisphere summer (June/July) but that is a glancing observation.

The bulk (146 or 33.96%) relate to Supply Agreements (General) and the rest to NDAs.

70 Annexure: Lawgeex Legal Operations Report.

21



When evaluating size, 44.47 or 197 contracts are less than 5 pages and considered “short”, 

while 20.09% or 89 contracts are more than 20 pages and considered “long”.

ii. Lawgeex Report 2: Legal Policy – Mutual NDA71

This covers the period from 9 January 2022 to 9 January 2023 (note 1). The additional note 

about displaying data for 6 June 2021 to 9 January 2023 is to remind the user that no legal 

policy data is available before 6 June 2021. This has no bearing on the data under review.

The Top Deleted Concepts out of 96 NDAs are listed as 1) Marking of Confidential 

Information; 2) Indemnification and 3) IP Assignment.

71 Annexure: Lawgeex Legal Policy - Mutual NDA.

22



The Top Missing Concepts out of 96 NDAs are listed as: 1) Confidentiality of Relationship; 

2) No Future Obligations and 3) Standard of Care.

The Top Revised Concepts out of 96 NDAs are listed as 1) Return or Destruction of 

Confidential Information; 2) Governing Law and Venue; and 3) Term of Confidentiality.

23



iii. Lawgeex Report 3: Legal Policy – Receiving NDA72

This covers the period from 9 January 2022 to 9 January 2023 (note 1). The additional note 

about displaying data for 6 June 2021 to 9 January 2023 is to remind the user that no legal 

policy data is available before 6 June 2021. This has no bearing on the data under review.

The Top Deleted Concepts out of 93 NDAs are listed as 1) Indemnification; 2) IP 

Assignment and 3) Liquidated Damages.

The Top Missing Concepts out of 93 NDAs are listed as: 1) Affiliate Definition; 2) No Future 

Obligations and 3) Compelled Disclosure.

72 Annexure: Lawgeex Legal Policy - Receiving NDA Report.

24



The Top Revised Concepts out of 93 NDAs are listed as 1) Governing Law and Venue; 2) 

Return or Destruction of Confidential Information; and 3) Permitted Disclosure.

iv. Lawgeex Report 4: Legal Policy – Supply Agreement (General)73

This covers the period from 9 January 2022 to 9 January 2023 (note 1). The additional note 

about displaying data for 6 June 2021 to 9 January 2023 is to remind the user that no legal 

policy data is available before 6 June 2021. This has no bearing on the data under review.

73 Annexure: Lawgeex Legal Policy - Supply Agreement Report note 1.

25



The Top Deleted Concepts out of 146 Supply Agreement (General) are listed as: 1) Financial 

Guarantees; 2) Governing Law and Venue and 3) Most Favoured Nation.

The Top Missing Concepts out of 146 Supply Agreement (General) are listed as: 1) Use 

Restrictions; 2) Affiliate Definition and 3) Export Limitations.

26



The Top Revised Concepts out of 146 Supply Agreement (General) are listed as: 1) 

Warranty; 2) Force Majeure; and 3) Limitation of Liability.

The dashboards as supplied by the Lawgeex platform provide a comprehensive overview of 

workload and details to deleted concepts, missing concepts and revised concepts. The

example contract in the ANNEXURES gives a practical example of the contract before the

upload and the contract as received from the platform. Deleted concepts are struck through, 

missing concepts are added, and revised concepts have text added to the original version. The

user has a clear view of all the amendments made by the AI platform, with explanatory notes

where necessary.

(b) Target Group Data Analysis Report – Legal Counsel or Contract Reviewers

A baseline was determined by polling all employees worldwide that have legal training and 

would have responsibility for legal contract review. The questionnaire74 went to 32

employees across 13 locations and had a response rate of 41%. Although Lawgeex has been

implemented for more than 3 years within the Endress+Hauser Group, the respondents were 

asked about their experiences before the introduction of Lawgeex. The questionnaire had

74 Annexure: Endress+Hauser Data Questionnaire: Legal Counsel or Contract Reviewer.

27



target questions and predefined responses like the available data from the Lawgeex 

dashboard, to determine the baseline on contract clauses but also contained effectiveness 

questions to determine the time and cost of the legal counsels and contract reviewers. The 

effectiveness results were as follows: Question 1) Number of contracts per week: 31% of 

respondents receive 2-5 contracts per week; 23% receive 5-8 contracts per week and 23% 

receive 8-10 contracts per week. Only 15% of respondents receive more than 10 contracts per 

week. Question 2) The average time spent per 10 contract pages: 46% of respondents spent

60-120 minutes per 10 contract pages; 38% spent 30-60 minutes and 15% spent 15-30

minutes to review 10 contract pages. Question 3) Number of NDAs per week: 15% of

respondents receive 8-10 NDAs per week to review; 54% of respondents receive 2-5 NDAs

to review; 31% receive less than 2 NDAs per week. Question 4) The average time spent per 3

NDA pages: 23% spent 45-30 minutes; 31% spent 15-30 minutes; 15% spent 10-15 minutes 

and 31% spent less than 10 minutes to review a 3-page NDA.

Almost 70% of the respondents answered negatively about their knowledge of their standard

production cost rate. Each employee in Endress+Hauser has a known “standard production 

cost” rate that reflects the accounting cost of everyone. It is more than the traditional cost-to- 

company as it accounts for building overheads as well as several other costs.

To compare the responses of the respondents, specific questions were posed with the 

Lawgeex dashboard data and had the following responses:

28



Top Negotiated Terms: 1) Liability clauses; 2) Delivery penalties; 3) Warranty / Guarantee 

duration.

Top Deleted Concepts: 1) Uncapped product liability; 2) Consequential damages; 3) 

Warranty / Guarantee period exceeding 36 months.

29

Respondents - Top Deleted Concepts
0 2 4 6 8 10 12

1. Uncapped product liability

2. Consequential damages due to product failure

3. Warranty guarantee period exceeding 36 months

4. Credit terms for non / new customers

5. Liability for Confidentiality breaches

6. Other (please expand)



Top Missing or Added Terms: 1) Limitation of liability; 2) Payment Milestones; 3) Force 

majeure clauses.

Top Revised Terms: 1) Uncapped liability to capped liability; 2) [Tie between three options]

Penalty application against total value versus outstanding items; Contract termination clauses 

and Transfer of ownership and risk.

30

Respondents - Top Missing or Added Terms
0 2 4 6 8 10 12

4. Limitation of liability

2. Payment milestones

1. Force majeure / Pandemic clauses

5. Compliance with International Law

6. Other (please expand)

3. Forward Exchange Contracts

Respondents - Top Revised Terms
0 2 4 6 8 10 12

1. Uncapped liability to capped liability.

4. Penalty application against total order versus
outstanding items.

7. Contract termination clauses.

8. Transfer of ownership and risk.

6. Incoterms2010 – FCA; DAP; CIF; EWX; etc.

3. Offset against payment versus claims process.

2. Fixed delivery duration versus fixed delivery date.

5. Settlement discount for early payments.

9. Other (please expand)



To gauge the respondents’ natural progression or regression from known legal arenas, three

additional questions not covered by the Lawgeex dashboard were asked 1) Which governing

law would you accept, other than your own country’s law? 2) Which governing law would 

you not accept, irrespective of the business potential? And 3) Which governing law would 

you propose as an alternative to your own country’s law? The results were as follows:

Alternative Governing Law: 1) Swiss; 2) England and Wales; 3) United States

31

Respondents - Alternative Acceptable Governing Law
0 1 2 3 4 5 6 7 8

4. Switzerland

1. England and Wales

2. United States

3. Singapore

5. Other (please expand)



Unacceptable Governing Law: 1) Russian Federation; 2) Any African country; 3) Any non- 

English language country.

Alternative Language to English: 1) German; 2) [Tie between three options] French; Spanish 

and “local language”.

32

Respondents - Unacceptable Governing Law
0 2 4 6 8 10 12

2. Russian Federation

1. Any African country

4. Any non-English language country

3. Any Latin American Country

5. No response

6. Other (please expand)

Respondents - Alternative Language to English
0 1 2 3 4 5 6 7 8

1. German

2. French

4. Spanish

5. No response

6. Other (please expand)

3. Chinese



(c) Target Group Data Analysis Report – Contract Uploaders

Following the baseline determination, a questionnaire75 was sent to all employees who had

uploaded a contract in the year under review. These are not necessarily staff with formal legal

qualifications but are perhaps dealing with a customer contract. A total of 74 employees 

received the questionnaire across 39 countries. A survey response of 53% was recorded. Like

the legal counsel questionnaire, effectiveness questions were also asked of this larger group.

The effectiveness questions had the following responses: Question 1) Who reviewed 

contracts before Lawgeex? 60% of respondents reported they reviewed contracts while 13%

reported that the entity Managing Director reviewed contracts. 13% reported external lawyers

were tasked to review contracts. Question 2) Time spent on 10 pages contract review: 48% of

respondents spent 30-60 minutes reviewing a 10-page contract while 24% of respondents 

spent 60-120 minutes and 19% of respondents spent 15-30 minutes reviewing a 10-page 

contract. [The remaining 10% was split between other options.] Question 3) Who reviewed

NDAs before Lawgeex? 63% of respondents reported that they reviewed NDA while 11%

reported that the Managing Director of the entity reviewed NDAs and 11% reported that

Heads of Department reviewed NDAs. 9% of respondents reported that external lawyers did

the review of NDAs. [The remaining 7% were split between other options.] Question 4) Time

spent on 10-page NDA review. 24% of respondents spent 30-45 minutes on a 10-page NDA 

while 22% of respondents spent 15-30 minutes on the same. 19% of respondents spent less 

than 10 minutes but 16% spent more than 45 minutes. 14% of respondents spent 10-15

minutes reviewing a 10-page NDA and the remaining 5% had external reviewers of NDAs.

Question 4) Turnaround time if the review was done by others: 36% of respondents reported

that the review was returned the same week and 33% reported that the review was received in

2-3 days. 19% reported that the review took more than 1 week. Only 8% reported that the 

review was received in 24 hours and only 3% (a single response) was that the review was 

received the same day.

Also here, the positive response to the cost question was equally low under this group of

respondents. 59% did not know their respective standard production cost. Also, 54% of 

respondents stated that they had in-house colleagues dedicated to contract or NDA review.

75 Annexure: Endress+Hauser Data Questionnaire: Contract Uploader.

33



(d) Data Analysis Discussion and Deductions

The Lawgeex platform effectiveness measures in terms of time, cost and quality are reflected 

in the dashboard data. The average turnaround time is 1 day 1 hour and 49 minutes of human 

working time. Even the best reviewers would work 1 hour per 10 pages and then receive 5

contracts with 20 pages which would generate a workload of 10 hours. However, with the

time that AI spent reviewing the contract, the human uploader or legal person can spend time 

on other activities. That is where the benefit lies. Some respondents stated that after 

implementing Lawgeex, the organization’s turnaround times for specific contract reviews 

decreased by up to two days.76

As to cost, the annual licence fee of $85,000 for the entire Endress+Hauser Group of 

companies, would yield a total of 460 manhours, considering that the average rate of

respondents was given as €150/hour or about $163/hour. This is slightly less than 3 months’

worth of manhours. There can be no doubt that from a cost perspective, the AI platform 

outperforms the human counterpart by a large margin.

The accuracy part of the platform yielded results that required insight. In the supply 

agreement review, the top deleted clauses by the platform were Financial Guarantees and

Governing Law, whereas the human counterparts stated uncapped product liability and

consequential damages. It is my opinion that these clauses would be amended by their human

counterparts instead of outrightly deleted. When the top revised concepts are evaluated, the 

AI platform gave warranty and force majeure, whereas the human counterparts gave 

uncapped product liability again and penalties on part order instead of the entire order. In the 

NDA review, the AI platform gave indemnification as a top deleted concept and governing

law and return or destruction of confidential information as the top revised concepts. None of

these clauses featured in the human counterpart responses. It was clear that the human 

contract reviewers placed a fair amount of emphasis on operational contract and agreement 

conditions and less emphasis on purely legal aspects. The legally trained contract reviewers 

did identify some of the legal aspects but the results do not show a consistent identification of

such clauses. As a Euro-centric organisation, the default fallback of many of the human

reviewers was to known legal positions. This included the German language in case of a non-

English language contract and Swiss governing law in the case of governing law alternatives. 

The response regarding governing law originating in Africa was the option selection to

76 Op cit note 43.

34



disregard any African country governing law. The same for any Latin American country too.

This demonstrated a lack of understanding of other legal systems and would need better

guidance from legal specialists if the business potential of these regions were to be tapped. 

(The exclusion of the governing law of the Russian Federation was not surprising, given the

Russian-Ukraine conflict situation of 2022 and 2023.) The specialisation of the human

reviewers demonstrated an understanding of the business environment, and the contract 

review was done in a similar light. Project management expertise and transactional selling

skills set the tone for the contractual conditions but the NDA reviewers missed some key

points from a legal perspective. The understanding of the functioning of the NDA clauses

was present, but the limitation of risk and consequences of actions was not always clear. The

Lawgeex solution greatly improved the accuracy of the reviewed contracts and agreements by 

highlighting clauses that were either lower in priority or even missed by human reviewers. 

The revised phrasing of the Lawgeex platform mark-ups helped to add clarity and the 

playbook structure also increased the accuracy and consistency of outputs.77

(e) Data Analysis Conclusion

The Lawgeex platform proved to be an effective tool in the parameters of time, cost and

quality, for the automated contract review when compared to human reviewers. The time and

cost aspects were to a certain extent anticipated, but the quality level surpassed expectations.

How important legal clauses and concepts were brought to the fore and marked up for clarity,

was far better than expected. The playbook model used by the platform proves that it is a 

sustainable model that consistently provides good results.

VI RECOMMENDATIONS

(a) General

The research that was done for this report would require additional development to determine 

its impact on the wider legal fraternity. The narrow focus may have diluted some findings. A

future research project could consider different AI service providers as well as different types

of firms, to support the findings. The results from the empirical study would need to be 

replicated in other areas with similar parameters to validate the findings.

77 Op cit note 43.

35



The introduction of an AI contract review platform into an organisation would require a clear

mandate from the business partner as well as buy-in from all the users. A clear “why?”

statement needs to be formulated to ensure that the tool is used as intended to deliver the 

results as expected. The formulation of a playbook or similar guide for the AI to follow is a

critical success factor. In this step, the legal counsel of the business entity should provide

clear direction on the requirements of the AI platform. The question of whether AI contract

review is beneficial, I believe has been answered: the time improvement, the cost reduction 

and the accuracy of incompatible clause identification, are consistently better than a human 

contract reviewer.

(b) Recommendations Specific to the Target Group

Several reviewers stated that the Lawgeex platform is insufficient in their business 

environment purely due to language. The inability of the platform to work in languages other

than English is a major drawback. Considering the global presence and regional focus of the

Endress+Hauser Group, multi-language capabilities are a necessity rather than a simple 

feature. Even in some English-language countries, a requirement exists to be able to review

contracts in other languages. In analysing the output of the AI review platform, the ability of

machine learning seems to be lacking. Nowhere can the AI platform “be taught” that rejected

or amended clauses are incorrect. A case in point is that all contracts sent from South Africa,

have the governing law and arbitration clauses replaced with Swiss law and Swiss arbitration. 

Yet nowhere can the review give feedback to the platform about removed mark-ups.

Internally towards Endress+Hauser, it was concerning that such a low number of associates 

were informed of their cost (standard production cost) towards the company. A few reviewers

also commented that the response time of Lawgeex was too slow for their business

requirements. I do not agree with such comments since it would be more beneficial to 

manage customer expectations than to rush contract review. Several respondents lamented

Lawgeex mark-ups as “too much” and that such mark-ups require additional, specialised time

to resolve. Whilst the latter part is true and detail mark-ups take additional time from legally

trained specialists, the AI platform demonstrated the clauses are missed by the human 

counterparts that are not in alignment with the legal playbook. The consistency demonstrated

by the AI to find and mark up the inconsistent clauses outweighs the additional specialist

reviewer time requirements. If the clause's markup is incorrect, however, the feedback needs 

to go the machine learning route to optimise the AI platform. There is a gap in my opinion.

36



Lastly, there appears to be a gap in linking the contract size to the risk appetite of the 

business entity. The AI platform reviews the contract or agreement without consideration of 

the contract value. Much of the risks and liabilities are intricately linked to the contract value.

Yet the AI platform deals with all the contracts the same way. In the Endress+Hauser

business environment, an R200,000 or €10,000 contract and an R10 million or €500,000 will 

attract a decidedly different risk appetite. The AI platform review of the contract does not

capture this in my opinion. A business decision can decide to accept all the penalties and

liability clauses on a low-value contract, in the view to be expedient and secure the business. 

Whereas a large contract will require far more due diligence and risk assessment before the 

contract is signed and concluded.

VII CONCLUSION

The advent of high technology for big data issues will have an impact on the next generation

of lawyers or legal practitioners. Gone are the days of spending hours reviewing legal

contracts in a commercial environment. The new lawyer needs to be adept in the output of

technology for him or her to spend quality (billable) hours in front of their client. Associates 

and paralegals alike must become proficient in the adept in the use of technology for him or

her to become a high-performance individual. There is an inherent risk that the stagnant,

traditional legal practitioner work will become obsolete in the Fourth Industrial Revolution 

era. A significant case was made that data processing and automated contract review can 

enhance the efficiency of mundane legal practice work, while at the same time reducing costs 

and increasing the quality of the work. There is also an opportunity for specialist legal 

practitioners to step forward to assist companies in defining their legal playbook for use in 

automated contract review platforms. An identified risk of AI is the programme background, 

including bias or preference, will impact the quality of the output, thus rendering the advice 

potentially skewed. Thus, the lawyer of the future must be able to translate legal principles 

and requirements into a functioning set of rules for the AI platform to compare. The platform 

is ultimately only as good as the background programming. The high-performance law firm 

of tomorrow will be the one that uses technology tools to enhance its client offering.

Word Count: 10,528

37



VIII BIBLIOGRAPHY

Adams, Nathan-Ross “South African Company Law in the Fourth Industrial Revolution:
Does Artificial Intelligence Create a Need for Legal Reform?” (unpublished LLM 
thesis, University of the Western Cape, 2021), available at
https://doi.org/10.2139/ssrn.4052285, accessed on 23 January 2023.

Anon Introduction to AI Contract Review (2021) Lawgeex.

Anon “Big-Firm Intelligence for Small Law Firms.” LexisNexis 2018, available at
https://www.lexisnexis.com.au/en/insights-and-analysis/research-and-
whitepapers/2018/the-rise-of-legal-ai-big-firm-intelligence-for-small-law-firms,
accessed on 13 October 2022.

Anon “Lawyer vs Artificial Intelligence: A Legal Revolution.” LexisNexis 2018, available at
https://www.lexisnexis.com.au/en/insights-and-analysis/practice- intelligence/2018/
Lawyer-vs-AI-A-legal-revolution, accessed on 13 October 2022.

Anon “Six Ways Artificial Intelligence (AI) Is Transforming Law Practice.” LexisNexis 18
September 2020, available at https://www.lexisnexis.co.za/evolution-of-legal-
research/six-ways-artificial-intelligence-ai-is-transforming-law-practice, accessed on 13
October 2022.

Anon “How Will Artificial Intelligence Impact the Legal Profession?” Exigent (blog) 17
March 2020, available at https://www.exigent-group.com/blog/how-will-artificial-
intelligence-impact-the-legal-profession/, accessed on 13 October 2022.

Anon “LAWGEEX GRANTED PIONEERING PATENT ON CONTEXTUAL AI ENGINE
FOR CONTRACT ANALYSIS.” Lawgeex 24 February 2021, available at
https://www.lawgeex.com/pr/lawgeex-granted-pioneering-patent-on-contextual-ai-
engine-for-contract-analysis/, accessed on 8 November 2022.

Anon “About us: The Endress+Hauser Group” available at
https://www.endress.com/en/endress-hauser-group, accessed on 11 January 2023.

Anon “Artificial Intelligence Law | AI Law.” Michalsons (undated), available at
https://www.michalsons.com/focus-areas/artificial-intelligence-law, accessed on 23
January 2023.

Anon “What Is Industry 4.0 and the Fourth Industrial Revolution?” McKinsey 17 August
2022, available at https://www.mckinsey.com/featured-insights/mckinsey-
explainers/what-are-industry-4-0-the-fourth-industrial-revolution-and-4ir, accessed on
24 January 2023.

Anon “What Is the History of Artificial Intelligence (AI)?” Tableau, (undated), available at
https://www.tableau.com/data-insights/ai/history, accessed on 6 August 2023.

Buiten, Miriam C “Towards Intelligent Regulation of Artificial Intelligence.” European
Journal of Risk Regulation 10 no. 1 (March 2019): 41–59, available at
https://doi.org/10.1017/err.2019.8, accessed 8 June 2023.

Deamer, Kacey “What the First Driverless Car Fatality Means for Self-Driving Tech.”
Scientific American 1 July 2016, available on

38



https://www.scientificamerican.com/article/what-the-first-driverless-car-fatality-means-
for-self-driving-tech/, accessed on 6 August 2023.

Dormehl, Luke “Why AI Will Never Rule the World,” Digital Trends, 25 September 2022,
available at https://www.digitaltrends.com/computing/why-ai-will-never-rule-the-world/,
accessed on 24 January 2023.

Fischer, Christian & Cohen, Ashleigh The Total Economic Impact™ of Lawgeex Forrester
May 2021, available at https://www.lawgeex.com/wp-
content/uploads/2021/10/May2021-Forrester_TEI_Report-LawGeex.pdf, accessed on 13
October 2022.

Galeon, Dom “Bill Gates Says We Shouldn’t Panic about Artificial Intelligence.” Futurism
26 September 2017, available at https://futurism.com/bill-gates-says-we-shouldnt-panic-
about-artificial-intelligence, accessed on 23 January 2023.

Greenstein, Stanley “Preserving the Rule of Law in the Era of Artificial Intelligence (AI).”
Artificial Intelligence and Law 30, no. 3 (1 September 2022): 291–323, available at
https://doi.org/10.1007/s10506-021-09294-4, accessed on 6 August 2023.

Haggerty, Jarrod “Disruptive Technology in the Legal Profession.” Deloitte United Kingdom,
2017, available at https://www2.deloitte.com/uk/en/pages/financial-advisory/articles/the-
case-for-disruptive-technology-in-the-legal-profession.html, accessed 23 January 2023.

Heinlein Robert A “Quotable Quote” goodreads (undated), available at
https://www.goodreads.com/quotes/12051-a-human-being-should-be-able-to-change-a- 
diaper, accessed 23 January 2023.

Kaminski, Margot E “Regulating the Risks of AI.” University of Colorado 19 August 2022,
available at https://doi.org/10.2139/ssrn.4195066, accessed on 6 August 2023.

Lant, Karla “Elon Musk: Competition for AI Superiority at National Level Will Be the ‘Most
Likely Cause of WW3.’” Futurism 4 September 2017, available at
https://futurism.com/elon-musk-competition-for-ai-superiority-at-national-level-will-be-
the-most-likely-cause-of-ww3, accessed on 23 January 2023.

Lant, Karla “Experts Weigh in on AI and the Singularity.” Futurism 7 June 2017, available at
https://futurism.com/experts-weigh-in-on-ai-and-the-singularity, accessed on 23 January
2023.

Marwala, Tshilidzi Closing the Gap (2020) Pan Macmillan South Africa.

Mik, Eliza “AI in Negotiating and Entering into Contracts.” AI in Private Law 1 June 2021.
Available at https://papers.ssrn.com/abstract=3873071, accessed on 6 August 2023.

Mollick, Ethan “All My Classes Suddenly Became AI Classes.” One Useful Thing (And Also
Some Other Things), 17 January 2023, available at
https://oneusefulthing.substack.com/p/all-my-classes-suddenly-became-ai, accessed on
23 January 2023.

Parapontis, Ioannis “AI in Law: Robot Lawyer in First Ever Court Case. Lawyers to Become
Obsolete in the Future? | LinkedIn.” Artificial Owl 8 January 2023 available at
https://www.linkedin.com/pulse/ai-law-robot-lawyer-first-ever-court-case-lawyers-

39



ioannis/?trackingId=PKSJgCxNSHWkgYZwJh98Ow%3D%3D, accessed on 9 January
2023.

Pfeifer, Jeff “LexisNexis: Legal AI and Our Data-Driven Future.” Artificial Lawyer (blog) 11
August 2017, available at https://www.artificiallawyer.com/2017/08/11/lexisnexis-legal-
ai-and-our-data-driven-future/, accessed on 23 January 2023.

Reedy, Christianna “Kurzweil Claims That the Singularity Will Happen by 2045.” Futurism
5 October 2017, available at https://futurism.com/kurzweil-claims-that-the-singularity-
will-happen-by-2045, accessed on 23 January 2023.

Russell, Stuart, Daniel Dewey, and Max Tegmark “Research Priorities for Robust and
Beneficial Artificial Intelligence.” AI Magazine 36 no. 4 (31 December 2015): 105–14, 
available at https://doi.org/10.1609/aimag.v36i4.2577, accessed on 13 October 2022.

Schwab, Klaus The Fourth Industrial Revolution (2017) Penguin UK.

Stepka, Matthew “Law Bots: How AI Is Reshaping the Legal Profession.” Business Law
Today 21 February 2022, available at https://businesslawtoday.org/2022/02/how-ai-is-
reshaping-legal-profession/, accessed 13 October 2022.

Surden, Harry “Artificial Intelligence and Law: An Overview.” Georgia State University Law
Review Volume 35 Issue 4 Summer 2019 (1 June 2019), available at https://
papers.ssrn.com/abstract=3411869, accessed on 23 January 2023.

Tegmark, Max “Hawking Reddit AMA on AI.” Future of Life Institute (blog), 12 October
2015, available at https://futureoflife.org/ai/hawking-reddit-ama-on-ai/, accessed on 13
October 2022.

Vanni, Rachel “How Artificial Intelligence Is Transforming the Legal Profession,” Kira 8
May 2020, available at https://kirasystems.com/learn/how-artificial-intelligence-is-
transforming-the-legal-profession/, accessed on 13 October 2022.

Wolfram, Stephen “Dropping In on Gottfried Leibniz.” Stephen Wolfram Writings 14 May 
2013, available at https://writings.stephenwolfram.com/2013/05/dropping-in-on-gottfried- 
leibniz/, accessed 23 January 2023.

40



IX ANNEXURES

(a) Lawgeex Document Examples

i Contract Draft – Original

ii Contract Draft – Lawgeex Reviewed

(b) Questionnaire Formats and Analysis

i Endress+Hauser Legal Counsel Questionnaire and Results

ii Endress+Hauser Contract Uploaders Questionnaire and Results

(c) Lawgeex Dashboard Reports

i Report 1: Legal Operations

ii Report 2: Legal Policy – Mutual NDA

iii Report 3: Legal Policy – Receiving NDA

iv Report 4: Legal Policy – Supply Agreement (General)

(d) Respondents' Reports and Graphs

i Graph Respondents - Top Negotiated Terms

ii Graph Respondents - Top Deleted Concepts

iii Graph Respondents - Top Missing Terms

iv Graph Respondents - Top Revised Terms

v Graph Respondents – Alternative Governing Law

vi Graph Respondents - Unaccepted Governing Law

vii Graph Respondents - Alternative Language

41



MINOR SERVICES AGREEMENT

This Minor Services Agreement (the "Agreement") is dated 18/07/2022 (the "Effective Date") and made 

between:

XXXX a company incorporated under the law of Mozambique, with registration number XXXX, on page 

57, from book C-45, on March 10, 2006, with tax registration number XXXX, with principal place of 

business at XXXX, Maputo- Mozambique, hereinafter referred to as the "Company"

and

[… …], with its registered office at [… …], (hereinafter the "Contractor"),

Each the Company and the Contractor referred to as a "Party" or, collectively, as the "Parties", who have 
agreed to enter into the Agreement on the following terms and conditions:

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of the 1/21
signed by and between XXXX and Endress Hauser.

i04100186
Text Box
ANNEXURE (a) i: CONTRACT DRAFT - ORIGINAL



1. Subject to instructions received by the Contractor from the Company from time to time, the

Contractor shall perform the Services for the Company, at the times and in the manner, described in 

Schedule A (the "Services").

2. The following Schedules attached to this Agreement shall constitute an integral part hereof:

a) Schedule A (Scope of Services);

b) Schedule AI - (Comercial and Technical Proposal);

c) Schedule AII - (RG 0006);

d) Schedule AIII - (RG 0018);

e) Schedule B - (Table of Quantities and Prices);

f) Schedule C – (Measurement Criteria)

g) Schedule D – (insurance)

2.1. If there is any conflict or inconsistency between the provisions of this Agreement and its

Schedules, the provisions of this Agreement shall prevail to the extent necessary to eliminate any conflict

or inconsistency. In the event of any conflict between the Schedules, the order of such schedules listed in 

Clause 2 above shall be observed.

3. It is hereby agreed that the Company shall pay the Contractor the total amount of ZAR 359.304
(three hundred and fifty-nine thousand, three hundred and four rands) for the Services during the term 

of this Agreement, in accordance with the payment terms specified in Schedule B (Bill of quantities).

3.1 All payments shall be made via wire transfer to the Contractor's bank account as follows: [… …].

Bank:

Bank Address:

International Bank Code (IBAN):

Account Number:

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of the 2/21
signed by and between XXXX and Endress Hauser.



Account Holder

3.2 Payment shall be made to the Contractor within 30 days of receipt by the Company of an invoice in 

form and substance previously approved by the Company.  If the Company disputes any item in an 

invoice, the Company shall not be obliged to pay the disputed item until the Parties have resolved

such dispute in an amicable way or in accordance with Clause 20, as the case may be.

3.3 All or part of any payment under an invoice may be withheld by the Company, or set-off against the 

payments owing to the Contractor, on account of: (i) defective Services rendered by the Contractor;

(ii) all debts, damages and moneys due and unpaid from the Contractor to the Company; and (iii) any

amount that the Company is required to withhold or deduct by law or by order of any relevant 

authority. Unless Contractor provides the Company with evidence of a waiver from the appropriate 

Governmental Authorities, Company may, where required by applicable law and subject to any 

applicable tax treaties, withhold from the aggregate amount charged all applicable withholding taxes,

which Company shall remit to the appropriate Governmental Authorities. Upon Contractor’s request,

Company will provide evidence of such remittance. Company shall provide Contractor with all tax 

information and documentation with respect to the withholdings as required by law and in the time 

required by law. Contractor shall be solely responsible for obtaining a refund with respect to any 

amounts withheld pursuant to this Clause.

4 The term of this Agreement shall be valid for 03 (three)  months counted from the issuance of the first 

Service Order, ending at the end of the referred period, without prejudice of the provisions of this

Agreement regarding the deadline for payment of the invoices.

5 The Contractor shall:

5.1.1 exercise that degree of skill, diligence, prudence and foresight that would reasonably and 

ordinarily be expected from time to time from a skilled and experienced professional seeking in

good faith to comply with its contractual obligations, complying with all applicable legislation and

laws and engaged in the same type of undertaking, and under the same or similar circumstances

and conditions as those envisaged by this Agreement, and in accordance with the Company's 

environmental, health and safety standards and procedures contained in Schedule E (Company 

Procedure), and instructions, as communicated by the Company to the Contractor from time to 

time;

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of the 3/21
signed by and between XXXX and Endress Hauser.



5.1.2 allocate sufficient resources and personnel for the performance of the Services;

5.1.3 progress the Services promptly and diligently;

5.1.4 prior to the date on which performance of the Services commences, obtain and maintain all 

necessary licenses and consents, and comply with all relevant legislation, in relation to the

Services; and

6 The Company shall:

6.1.1 make payments to Contractor for the performance of the Services, in accordance with the terms

of this Agreement and Schedule C (Commercial Conditions);

6.1.2 without warranty or liability, provide the Contractor all information in its possession which, in the 

opinion of the Company, is reasonably required for the performance of the Services and pertains

to the Services; and

6.1.3 on all matters properly referred to it by the Contractor in connection with this Agreement, give 

its decision within the time periods specified in this Agreement, or within a reasonable period if

no such period is specified herein.

7 Any notice, consent or approval required or permitted to be given in connection with this Agreement 

shall be in writing, in the English language and be addressed to a Party at the address chosen by the

Party concerned below:

7.1.1 in the case of the Company to:

XXXXXXXX

XXXXXXXXX

Tete – Moatize - Moçambique

Att: XXXXX

Tel:  XXXXX

E-mail: XXXXX

7.1.2 in the case of the Contractor to:

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of the 4/21
signed by and between XXXX and Endress Hauser.



[___________]

Name:

Fax:

Telephone number:

E-mail:

Attn:

8 Either Party may terminate this Agreement for convenience by delivering a written notice to the other 

Party, and such termination will be effective from the time stated in the termination notice, or if no

such time is stated in that notice, at the time such notice is delivered to the other Party.

8.1.1 In addition, if any of the Parties commits a material breach of this Agreement (which shall include, 

with respect to the Contractor, any Contractor Event of Default and, with respect to the Company,

be limited to any Company Material Breach) and, upon receiving a Contractor Event of Default

Notice or, as applicable, Company Material Breach Notice from the non-breaching Party, such 

material breach is capable of remedy, and, the breaching Party fails to remedy such material 

breach within [30 days] of receipt thereof, the non-breaching Party shall be entitled to terminate 

this Agreement by further written notice to the breaching Party.

8.1.2 Termination for Contractor Default

8.1.2.1 The Contractor will be in default under this Agreement if any of the following events occurs (each,

a "Contractor Event of Default"):

8.1.2.1.1 dissolution, which includes the bankruptcy, insolvency, liquidation, amalgamation, 

reconstruction, reorganisation, administration, administrative or other receivership, winding

up or dissolution of the Contractor, and any equivalent or analogous proceedings howsoever

described and in whatever jurisdiction, and any step taken (including the presentation of a 

petition or the passing of a resolution) for or with a view to any of the foregoing (each an 

"Event of Insolvency") occurs and is continuing, with respect to the Contractor;

8.1.2.1.2 the Contractor commits a breach of the environmental health and safety plan or applicable

laws relating to the environment, health and safety;

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of the 5/21
signed by and between XXXX and Endress Hauser.



8.1.2.1.3 the Contractor makes a false or misleading representation, declaration or warranty, and, in 

the case of a false or misleading representation, declaration or warranty, that is capable of

being remedied, such misrepresentation is not remedied within the period set forth in Clause

8.1 above;

8.1.2.1.4 the Contractor abandons performance of the Services;

8.1.2.1.5 the Contractor assigns and/or novates its rights and/or obligations under this Agreement; 

8.1.2.1.6 the Contractor enters into a subcontract for any part of the Services without the prior written

consent of the Company;

8.1.2.1.7 the Contractor fails to effect, have in place or otherwise maintain, or provide evidence of, 

insurance as required under Clause 17 and Schedule F (Insurance and Worker´s

Compensation);

8.1.2.1.8 the Contractor does not progress the Services in a timely and diligent manner; or

8.1.2.1.9 the Contractor breaches Clauses 17.2.1 to Error! Reference source not found. inclusive. 

8.1.2.2 If any Contractor Event of Default occurs, the Company shall be entitled to terminate this

Agreement by written notice to the Contractor (each, a "Contractor Event of Default Notice") in 

accordance with Clause 8.1 above, without prejudice to any of its other rights and/or remedies.

8.1.3 Company Material Breach

8.1.3.1 The Company shall be in Material Breach of this Agreement if any of the following occurs (each a

"Company Material Breach"):

8.1.3.1.1 an Event of Insolvency occurs and is continuing with respect to the Company; or

8.1.3.1.2 the Company fails to make an undisputed payment that is due and payable within the 

payment period in accordance with this Agreement and upon receiving notice of such breach

from the Contractor, the Company fails to remedy such breach within 30 business days of its

receipt of such notice of such breach.

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of the 6/21
signed by and between XXXX and Endress Hauser.



8.1.3.2 The Contractor must give written notice to the Company of the occurrence of a Company Material

Breach (each, a "Company Material Breach Notice").

8.1.4 Automatic Termination for certain Contractor Events

It is hereby agreed that if the Contractor and/or its Affiliates and Subsidiaries ever engages forced labour,

child labour or any other labour conditions that oppress human rights and dignity, this Agreement and/or

any other contract entered into by the Contractor and/or its Subsidiaries and/or its Affiliates, whether 

directly or indirectly, with the Company, will be automatically terminated. In such case, termination of 

this Agreement will occur immediately and automatically, the Contractor will not be entitled to any prior

notice or cure or remedy rights or periods and such termination will be without prejudice to proper notice

to the competent authorities and any further applicable rights and/or remedies of the Company in relation 

thereto.

8.1.5 In the event of termination of this Agreement by either Party under Clause 8, the Company shall 

pay the Contractor the aggregate amount of all undisputed invoices which remain due/not unpaid

as of the relevant termination date in relation to Services performed by the Contractor hereunder.

The Company shall only be obliged to make payment of any amounts under this Clause 8.1.5 30 

days after the date on which Company has approved the relevant invoices submitted by the 

Contractor for payment on termination.

8.1.6 In the event of termination of this Agreement pursuant to Clause 8.1.1 with respect to a 

Contractor Event of Default or Company Material Breach, the Party in default or, as applicable, in 

material breach, shall pay the non-defaulting or, as applicable, non-breaching, Party liquidated 

damages in the aggregate amount of 10% of the total price payable under Clause 3 of this 

Agreement. The Parties hereby agree and declare that liquidated damages payable under this 

Agreement are a genuine pre-estimate of the non-defaulting or, as applicable, non-breaching,

Party's loss under the relevant circumstances, and will be a debt due and payable by the defaulting

or, as applicable, breaching, Party to the non-defaulting or, as applicable, non-breaching, Party. 

For the avoidance of doubt, the Contractor shall not be entitled to any compensation under this

Sub-Clause 8.1.6, if termination of this Agreement occurs under Sub-Clause 8.1.4.

8.1.7 Termination for Force Majeure

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of the 7/21
signed by and between XXXX and Endress Hauser.



If a Party is relieved from the performance of any obligation under Clause 18 for a continuous period of

[180] days or more; then the other Party may, at any time thereafter, by written notice to the other Party, 

terminate this Agreement with immediate effect.

8.1.7.1 Payment on Termination for Force Majeure

If this Agreement is terminated by any Party under Sub-Clause 8.1.7, the Contractor's sole compensation 

will be paid for direct costs reasonably and properly incurred as a result of any such termination.

9 Each Party must keep confidential and shall not, without the prior written consent of the other Party,

disclose to any Person:

9.1 the details of this Agreement, nor the details of any of the transactions or agreements contemplated

by this Agreement;

9.2 any information relating to the business, operations, affairs, assets and liabilities of the other Party or

any of the other Party's subsidiaries, holding companies or Affiliates;

9.3 any non-public, confidential, personal or proprietary information disclosed by one Party to the other, 

whether in writing or communicated orally or electronically (and including all information obtained

by a Party through visual inspection of the assets of the other Party and any document prepared by a

Party that contains, or otherwise reflects or is generated, by using any information specified in this 

Paragraph 9(9.3, irrespective of how such information came into such Party's possession, and 

irrespective of how such information is recorded or stored),

and, in each case, whether or not the relevant information is marked as "confidential", collectively, the 

"Confidential Information".

9.3.1 The Parties may disclose Confidential Information to:

9.3.1.1  their respective officers, directors, employees, agents and professional advisors who: 

9.3.1.1.1.1 have a need to know (and then only to the extent that each such person has a need to know); 

9.3.1.1.1.2 are aware that the Confidential Information must be kept confidential;

9.3.1.1.1.3 are aware of the disclosing Party's confidentiality undertaking hereunder; and

Minor Services Agreement DM 1043 Commissioning of Vibratory Fork Density Meters conditions of th