Smart Contracts AI Article

Artificial Intelligence can
dramatically improve contract
intelligence, reduce legal risks and
minimize dispute costs
CONFIDENTIAL
Transnational Dispute Management
by Shannon Copeland
October 2014

02
TMD Technology Article October 2014
Companies are facing increasing financial risk
due to their decreasing ability to manage the
“big data” associated with complex contracts.
Specifically, large complex contracts are inherently
opaque and difficult to interpret. Large contracts
never exist alone but are interconnected with
hundreds if not thousands of related contracts
that have a material collective influence on the
overall risk picture.
The average Fortune 1000 corporation has
between 20,000 and 40,000 agreements or
contracts1
governing their everyday activities.
Information related to a corporation’s large contract
ecosystem is exploding in volume and speed, which
is sometimes referred to as “Big Data”, requiring
risk managers to find new tools, other than email,
to interpret the data.

03
1. Goldman Sachs - 2001, 2011, Gartner - Enterprise Contract Management, 2010 D. Wilson
2. “Hope is not an effective risk mitigation technique.” - Navigant (March 2012) - Brian C. Fox, James G. Zack Jr.
3. Technology in this context largely means information technology or software.
Big data’s components
Volume:
How much data must be analyzed; this is often
connected in the commercial context to how far
back in time must data be searched to make a
decision today.
Velocity:
How quickly or “how real-time” must data be
analyzed
Variety:
Numbers, text, structured, unstructured, video,
voice
Veracity:
Quality of the data or uncertainty of the data
Examples of this “new risk” are now infamous.
Lehman Brothers and other global banks simply could not
see or at least manage the collective risk of their financial
contracts. A more relevant industry for this paper is the global
“mega-project” construction industry. In the graphic below
“Megaproject Risk Sources”2
the inability to manage complex
contracts in the global large construction project market and
their resulting disputes cause an estimated 20% erosion of
contract value or $375 billion in losses each year.
The key question is whether there is
something new that can meet this challenge now.
Thus, can technology3
…
1. play a new and reinvigorated role in preventing
and mitigating legal, contract, regulatory and
compliance risk?
2. help business leaders and lawyers resolve
complex disputes more efficiently and amicably?
3. improve the creation, maintenance and
sustainability of commercial relationships?
4. fill current gaps to become a foundational
component in a modern risk management
program?
5. is this technology proven, affordable, and ready
today?
The answer is YES to all five questions.
Design
Budget
Front-end
planning
Technical
Procurement
Staffing
Organizational
Regulatory
Site
Market
Contract
specifications
Relative frequency
Relativeseverity
20% /
$375b
Megaproject Risk Sources
Project
execution
Disputes/
claims
Contract
administration

04
Professional elites, big data
and email don’t mix
As the old saying goes, wisdom is knowing what you don’t
know but for many professions, this phrase provides only
cold comfort; corporate risk managers know the knowledge
exists but the challenge is finding and synthesizing it quickly
enough to matter. Lawyers’, risk managers’ and arbitrators’
roles require the synthesis of many data types across time
periods but that data are almost never contained in one place
and are never organized, pre-processed or connected in a
useful way. Compounding the challenge of this issue is the
reality that email has become the dominant vehicle for critical
communications and legal workflow, yet “off the shelf” email
packages are in no way optimized to help these professionals
manage their jobs. Thus email becomes the easy scapegoat: in
interviews and collaboration with dozens of risk managers and
lawyers around the world, THE problem, called email, emerges
quickly in the conversation; the refrain is as follows:
“I receive over 300 emails every day and I spend
most of my time searching for and filing emails in
a vain attempt to prioritize what time remains to
respond thoughtfully.”
It is not surprising that research suggests this is a problem for
all knowledge workers and particularly for the “professional
elite”. IDC4
, an international research company, found in
2001 and again in 2014 that approximately 74% of knowledge
workers’ time was spent essentially not playing their real (or
most important) role: analyzing data and supporting decisions.
Not only has this ratio remained largely unchanged for the last
13 years, corporate data is now growing 55% to 85%5
annually
and 90% of this data is unstructured6
, primarily text. If email
is not the lone antagonist, its inherent inefficiency and lack of
integration with critical systems and sources of information
bear a significant portion of the blame.
The response thus far:
new hires and key word search
Organizations have attempted a response to the data deluge
described above by using techniques, now digital, that would
have been found in any well-run company in the 1930’s
and 1940’s. Documents and emails are meticulously tagged
and filed in document management systems (DMS), more
professionals are hired with additional junior support to
synthesize “everything”, assess risk (e.g., legal, operational,
monetary valuation) and report to committees that are hastily
established to make decisions with seemingly better, more
timely information. Downstream in the risk cycle and in some
cases after the fact, litigators and arbitrators are largely left with
the corporate human victims of these “processes” along with
keyword search or at best tools with long names like “latent
semantic indexing”7
. The most advanced teams have worked
with consultants and even law firms to unpack the types of
analyses that should be taking place. See case study following
as well as research excerpts from McKinsey & Co. and Navigant
Consulting8
.
4. IDC (idc.com) (2001, 2014); 32% of time was spent gathering or formatting information; 34% of time was spent searching, 8% was spent recreating
information, 26% of time was spent analyzing information. (The Knowledge Quotient; Unlocking the Hidden Value of Information, June 2014).
5. William Blair Investor Report e-Discovery, March 2012
6. IDC: unstructured content accounts for 90% of all digital information
https://idc-community.com/groups/it_agenda/business-analytics-big data/unlocking_the_hidden_value_of_information
7. Latent Semantic Indexing: http://en.wikipedia.org/wiki/Latent_semantic_indexing
8. With permission from McKinsey & Co. and Navigant

05
The solution:
smart collaboration and
artificial intelligence9
software
9. Nature.com - Computer science: The learning machines - Using massive amounts of data to recognize photos and speech, deep-learning computers are taking a
big step towards true artificial intelligence. - Nicola Jones 08 January 2014
10. In this article the terms machine learning, software, technology, artificial intelligence are used interchangeably.
11. For example, global banks are using machine learning to monitor for fraud, money laundering and other financial crimes.
12. Popular modern and historical examples of widely used smart workflow platforms include: Lotus Notes, Microsoft SharePoint, Salesforce.com, Siebel Systems,
HighQ.com, Wrike.com.
The solution to the big data problem, described
previously, lies in the development of software,
tools, applications and interactive platforms that
enable corporations to “feed” all the data into an
“intelligent agent” or “brain” that can augment expert
analyses. The new technology uses concepts such
as: collaboration, linking multiple sources of data
across sectors, regions and departments, non-stop
data mining and early detection of issues or non-
compliance to reduce risks, costs and conflicts.
Although machine learning10
is not new, it is not widely
known or used. Machine learning promises to help companies
build applications, affordably, that enhance and/or automate
knowledge work. Particularly interesting, as we will see in the
case study below, are problems that involve rules, regulations,
contracts and governance11
.
The highlights of machine learning are:
1. Intelligent ‘agents’ that help corporations
“piece” things together and detect/predict risks
earlier, while ‘learning’ and therefore lead to
improved accuracy over time
Gains to the legal risk management process are maximized
when legal knowledge workers are provided access to these
machine learning /AI capabilities through collaboration
technologies. Collaboration technology, sometimes referred to
as smart work flow technology12
, is a suite of software tools and
functions that help teams manage complex tasks and processes.
The most sophisticated collaboration “environments” allow
teams to automate some tasks such as information or document
routing (i.e. workflow) and approvals because the portal,
environment or “room” has embedded knowledge or logic
that enables the next step of the process. Therefore, it can
easily be thought of as “anti-email” technology. Today, most
professionals collaborate with their colleagues on projects large
and small through email. As described above, email’s simple
approach of one list one place has become too simple. Instead,
collaboration software is designed to match how the work is
performed, with some improvements. Salesforce.com is one of
the best relatively recent examples of a tool that has made the
sales or “customer relationship” process become more efficient
organizing, automating and extracting data from knowledge
work that had previously been under-optimized in email alone.
The example to the left is how a global energy equipment
manufacturer organized its contract and legal risk management
processes into a tool that avoided email by design. All
documents and emails from external parties were triaged in this
tool. Roles were established and checklists were embedded to
ensure the team was coordinated globally.

06
2. New technologies reduce legal risks and
costs by providing better and earlier detection and
allowing more time for legal and risk practitioners
Recently, while working with the CEO of a global
transportation systems company, the CEO and his legal contract
risk team developed the following “problem definition”:
… “the complexity of today’s infrastructure
projects demands world-class management at
every level in every role. The challenge is to identify,
develop, retain and deploy world-class expertise…
cost effectively. This means we have to have the
right expert at the right time at the right intensity
with the right information.”
Thus the technology tools described here, properly
implemented, work together to reduce risk by providing better
and earlier data and by providing (returning) more time to the
legal and risk practitioners to analyze the data.
3. Technology can detect weak areas early and
flag potential conflicts
Risk managers inside of construction and infrastructure players
and large banks alike know well the weak signals they must
detect in order to prevent material risks. Machine learning
systems can be ‘taught’ to look for these signals around the
clock. Examples include the ability to detect that a vendor or
contractor is slowing down their response time on critical issues
over the course of a project; or identifying as early as possible
delays in shipments, increases in prices, scope additions, turn
over in personnel, or the fact that documents are missing can
provide actionable more accurate data that is difficult and costly
to mine from email text. More subtle changes can be detected
as well, such as a change in word choice within emails. Risks,
particularly fraud, can often been disguised in text by the use of
arbitrary words that don’t fit the context around them. Machine
learning tools can memorize all words, over time, associated
with risk and can be taught to flag words that “aren’t known”
or defined. Global banks are deploying these tools to prevent
financial crimes.
These tools can eventually be used to detect how a party is
interpreting a particularly vague clause in a contract or perhaps
how a party views their own exposure to risk. These systems
need only be taught to detect the patterns that would only
emerge under any number of interpretations.

07
4. Using smart search engines, software can
automatically create risk registries, heat maps and
compile comprehensive files corresponding to each
risk category
For decades corporate lawyers and risk managers have suffered
from the same mistakes (i.e. lack of awareness) on similar
initiatives year after year from construction contracts to
derivatives instruments. Large infrastructure projects fail to
filter out sub-contractors that are ill prepared for the confetti
of data and change orders and global banks fail to detect fraud,
money laundering or the systemic risk of complex financial
instruments.
Similarly, as teams define or recognize issues in a risk register,
the registers can automatically mine data across the corporation
(i.e., federated searching) to ensure that each risk issue “file”
contains all relevant emails, documents and data.
These new data allow risk functions to perform the tasks that
consultants have long recommended like those contained in the
charts 5 and 6 below. For example, both McKinsey & Company
and Navigant Consulting recommend the use of risk registers
to produce “heat maps”. Risk registers are notoriously difficult
to define across even a single function and to populate or use
in a meaningful way. The tools described here can be used to
automatically populate a risk register and heat map once the
necessary standards are in place.
Machine learning13
enabled search engines can be used to
perform “graph searches” from within these collaboration
“rooms” to help experts mine historical and emerging data
repositories (i.e. email); these search engines, properly designed
can perform federated searches across “pre-indexed” content
to yield returns at speeds that consumers have come to expect
from public search engines like Google and Bing. Faster
corporate or enterprise search, allows risk experts to identify
patterns manually and then teach these concepts to the “AI
brain” and improve the corporation’s ontology. This allows for
an ever increasing set of “automated awareness” where experts
are asked to simply exclude the ever decreasing irrelevant data
and focus their time on the information that matters.
Naturally, all of these facts can be used to the benefit of the
company to negotiate better results in the fewer formal disputes
that will invariably emerge. Mr. Abrahamson’s vision of more
records, should now be affordable:
“
A party to a dispute will learn three
lessons (often too late): the importance
of records, the importance of records
and the importance of records.
”Max W Abrahamson
Engineering Law and the I.C.E. Contract
5. These new technologies return thousands
of hours to legal and risk professionals to better
synthesize the organized data
Collaboration tools and AI search engines can return thousands
of hours of time to a corporation’s ability to analyze risk.
Collaboration tools or “smart work flow” organizes content and
teams into processes that purposefully avoid email’s foibles.
Smart work-flow can also include automated rules that allow
tasks and sub-tasks to be assigned to internal and external
team members. Checklists or embedded knowledge tools can
be integrated into the workflow to enable delegation of tasks
to less experienced professionals, relieving further the burden
of non-analytical tasks on higher level experts. For example:
risk register items can be delegated based on level of monetary
exposure and complexity.
13. Nature.com - Computer science: The learning machines - Using massive amounts of data to recognize photos and speech, deep-learning computers are taking
a big step towards true artificial intelligence. - Nicola Jones 08 January 2014

08
Pitfalls
As corporations implement a standard, technology enabled
approach to risk management that takes advantage of
structured and unstructured data, the day of making only new
mistakes may have finally arrived. Elegant technology that is
easy to adopt and has the impact desired is hard to achieve.
Many argue, for example, that Apple’s recent improvement
from near financial failure to the most valuable company in
the world, depended on integrating existing and even old
technology into a platform that was elegantly designed (and
affordable). The same opportunities and risks exist for a
company that realizes it needs to look at technology and its
components that have either previously failed or have never
been tried.
In summary the challenges are obvious:
1) Humans required: Investments in this technology
should not be based on ROI calculations that depend on
reductions in human capital. Instead, professionals that have
been involved in these efforts find they are now better able
to perform their duties at the level they expect due to this
technology. This allows the team or corporation to compete
more effectively at the top and bottom line. (Please see the case
study included here).
2) Narrow the scope: The narrower a team can define
the problems they are trying to solve in the big data arena, the
easier and the more affordable it will be to create a something
that works. For example, a company may choose to start with
defining the top ten risks they wish to detect in a given area and
ensure that a risk register is automatically populated with these
risk items.
3) Fast first14
: Although there are elements or
components of this technology available, a complete solution is
not. The response should be to build a prototype quickly and
iterate the development. While many technology savvy readers
may disagree with this author’s opinion, building a prototype
quickly based on real- world problems, that is, with real-world
data, will help prove or disprove that efficiencies can be gained.
4) Programmers with commercial experience: Hiring
programmers that have experience with commercial or business
problems is critical; many AI programmers have spent their
careers working for governments or increasingly in research
functions like those contained in Google or IBM. While these
brilliant developers can and should play a role, there are an
increasing number of developers who have built practical and
more humble AI applications under the guidance of business
managers and risk managers.
14. Forbes: April, 2012, Steve Denning - “The Best-Kept Management Secret On The Planet: Agile”.

09
Case Study:
Large Construction Project Risk Mitigation
“Just working harder
will not produce better results.”
The following case study involves a global oil and gas product
manufacturing and services company. The company realized that
new, technology-assisted approaches were required to reduce risk,
particularly those risks that were encountered again and again on
similar projects. The traditional methods (e.g., more people, more
outside consultants) simply were not working.
Problem:
• Management has less than optimal visibility
and control of contractual risk.
• Inability to know:
» What are the sources (with a contract or regulation)
of risk at any given moment
» Inability to connect adverse events with a contract term
» Inability to assess risks in real-time sufficient to
mitigate or prevent formal disputes
» Inability to make permanent the lessons learned
from prior experience.

10
Act Differently
• “Upgrade” Contract
Administrator
• Dramatically improve
project documentation
• Interact more often with
construction attorneys
• Requests for assistance
• Detailed issue tracking
• Real time letter-writing
• Weekly issue calls
Track metrics
• At least 50% less
‘CA’ cost
• 11 folders/289 docs
• 26 “official”
• 10 issues
• 24 menos
• 35 conference calls
$16.0MM
($350,000)
Costs Net
benefit
$16.0MM
Other issues $0.3MM
Delay mitigation $1.0MM
Contractor issues $14.7MM
Inside and outside legal/risk team
Chart 2 : Process improvements in the legal/risk management area
Chart 3: Root “cause” of legal risk and savings.
Act Differently
• “Upgrade” Contract
Administrator
• Dramatically improve
project documentation
• Interact more often with
construction attorneys
• Requests for assistance
• Detailed issue tracking
• Real time letter-writing
• Weekly issue calls
Track metrics
• At least 50% less
‘CA’ cost
• 11 folders/289 docs
• 26 “official”
• 10 issues
• 24 menos
• 35 conference calls
$16.0MM
($350,000)
Costs Net
benefit
$16.0MM
($350,000)
Legal
Costs
Net
benefit
Other issues $0.3MM
Delay mitigation $1.0MM
Contractor issues $14.7MM
Inside and outside legal/risk team
orchestrated preparation of notice
letters that preserved contractual
positions and rights.
The new system, see Charts 2 and 3 below, introduced new
software, processes and metrics that were tracked to ensure
change was accomplished. Results were also observe and
recorded in detail, see Chart 1 “New Model”. The costs of
reducing or avoiding risk were reduced materially and the net
savings were monitored and measured/proved by category.
In Chart 2, weaknesses were identified in the traditional
contract risk management model; for example, the traditional
contract administration role was “unpacked” to identify
improvement area such as the ability to negotiate multiple
risk issues. Documentation was made more formal and more
detailed; discipline in other processes was improved by utilizing
a collaboration portal technology.
In Chart 3, below, issues and corresponding outcomes were
categorized and recorded to make it easier for future project
teams to learn from prior projects.
Phase 2: In an effort to automate portions of the smart
collaboration processes (i.e. smarter collaboration), Phase 2,
included plans to leverage machine learning to increasingly
automate the detection of contract risks. Thus the case study’s
conclusion was as follows.
$130m project – ($24m claim +$5m in legal fees) = 22% ROI erosion
Legaldisputecosts
Issue resolution period
E-Discovery Technology helps
but only after the factCorporate
data
growing
55% per
year*
Zero
risk
visibility
$3.0m $2.0m
Fact reconstruction Formal dispute resolution
Chart 1: Typical costs pattern for legal dispute
Previous similar projects had experienced the following pattern.
Preparation or facts reconstruction was consumed more time
than the formal dispute.

11
15. William Blair Investor Report e-Discovery, March 2012
Conclusion: Even minor investments in process
and technology can produce big returns in the
contract risk management area.
To ensure that a group’s legal/contract risk
mitigation is as efficient as possible, professionals
should embrace new technology. Because the
volume of corporate data is growing at 55%
annually15
, we have recognized that information
management can be a significant factor in managing
contract risk and winning disputes; at the request of
our clients, we are now guiding our clients through
“big data’s” advantages and disadvantages.
“New” technology has three primary components:
1. Advanced Collaboration: Advanced
collaboration technology and smart work flow
tools can help manage contract and compliance
risk on large sophisticated construction projects.
2. Advanced Search: Advanced search tools can
search ten million emails and documents in
less than one second even if the information
is distributed in separate databases around the
world.
3. Advanced Analytics and Awareness: Semantic/
artificial intelligence tools will allow commercial
teams to teach/institutionalize what to look for
ahead of time and then to automatically sense
changes in a claims, issues or situations that
the team needs to monitor. The system can
automatically, “24-hours-a-day”, identify new
emails and documents relevant to existing or
emergent issues and even to the specific clauses
and terms in the contract(s). A thoughtful,
affordable path should be developed to pursue
this vision. (see Chart 4).

12
As we have seen in our case study, there are immediate material
gains to be achieved by starting this path towards a more
modern approach to contract risk management. While these
immediate gains are achievable through sensible deployment of
rather pedestrian software, they begin to create the foundation
and towards ever increasing intelligence and assistance from
“self-learning” systems.
Chart 4 is an example of how a corporate function or
“department” could plan their development of an artificial
intelligence capability aimed at a narrow problem. The above is
for a large construction contract risk management “brain” that
could ultimately detect, flag and value risk issues in real-time.
First, it is helpful to think about degrees of accuracy required
(X axis) versus degrees of automation (Y axis)16
. “Leaping” to
very high levels of accuracy and automation is unreasonable
and unaffordable. For example, if a lawyer on the development
team wanted to automatically detect issues related to delays of
work described in the contract but required that the system flag
such delays without error (e.g., zero false positives, zero false
negatives), the development could not likely move forward.
Instead, such efforts are fueled by the realization that machine
learning can quickly achieve some level of automation with the
same or improved accuracy or vice versa.
In the example above, different technologies are deployed in
sequence over two to three years that help move the project
forward. Smart workflow is a relatively simple and inexpensive
first step, as described in the case study above, which helps
aggregate data, people and better processes. Semantic search
tools help speed up traditional searching while providing a
nascent ability to search using natural language17
like “give me
all the delays that occurred last month”. These early steps also
help the development team unpack the problem into more
achievable components. Human or expert tagging that requires
that professionals like risk managers and lawyers engage with
the emerging “AI brain” to provide feedback which accelerates
the system’s learning and accuracy. In short, human experts
that are engaged in the system make the system affordably
better faster.
Looking even further, in Chart 518
above, McKinsey identified
nearly 20 areas where corporations are failing to control
risk appropriately in the major construction infrastructure
market. As has been described in this paper, the primary
reason corporations continue to fall short is because email
is simply not enough and adding more experts is neither
possible nor affordable. New technology can help existing
managers perform better in the nine areas above. For example,
real-time risk related decisions can more accurately be made
if technology can help risk managers mine “big data” and
understand risk earlier. Risk registers can be automatically
populated with early issues and managed to a conclusion faster.
A global risk organization can collaborate globally in a standard
way to interpret risk and improve with each project.
In Chart 6 , Navigant similarly found, in 2012, that though
there are many “mathematically” proven risk management
techniques, that contract managers on large construction
projects, were not regularly deploying even the simplest
methods.
Expert Tagger
(Risk Manager
Tagging)
AutomationofContractRiskManagement
Accuracy
Client integration
Proof of concept development/tested
Development plan in place
Semantic email search ‘Outlook Plug-in’
Smart Workflow (Collaboration)
Auto Tagger
Contract auto-extraction
Awareness windows
NLP Queries
Chart 4: Artiﬁcial intelligence maturation path
NLP = natural language processing.
16. Artificial intelligence experts at (or formerly with) Vulcan Capital recommend this accuracy vs. automation framework.
17. Natural Language Processing (NLP)
18. McKinsey & Co.’s November 2013 study “A risk management approach to successful infrastructure projects”. (with permission). Also: IBM Watson’s potential
impact on the legal industry: http://www.abajournal.com/legalrebels/article/10_predictions_about_how_ibms_watson_will_impact/?utm_source=feeds&utm_
medium=rss&utm_campaign=site_rss_feeds

13
4
INTEGRATED
ENTERPRISE-RISK
MANAGEMENT
(ERM)
01
02
03
04
05
Insight and
risk transparency
Natural
ownership,
risk appetite
and strategy
Risk-related decisions
and processes
Risk
organization
and
governance
Risk culture and
performance
transformation
Risk culture: ensure soundness of risk
culture across entire organization (perform
culture diagnostic)
Risk norms: new risk norms need
to be embedded through various
corporate processes
and governance
Risk skill building:
implement a skill-enhancement
program for key roles
Risk archetypes: define ERM
mandate of the risk function
Risk organization: design risk
organization across entire
organization and ensure
appropriation of top management
Risk-function profile: establish
clear allocation of responsibilities
between risk taking and controlling units
Risk-related decisions: embed risk in business decision making
rather than pure compliance-oriented paper pushing
Risk optimization: embed in each major strategic decision
before launching/positive decision
Risk process: design and execute core business process-
es and operations on a risk-informed basis
Risk ownership: decide which
risks you own and which you don’t
Risk capacity: understand how
much risk you can take
Risk appetite: decide with how
much risk taking you feel
comfortable with
Risk strategy: decide on actions to
transform your risk profile, including
trade-offs with corresponding cost
Risk taxonomy: establish
common vocabulary for different risks
Risk register/risk heat map: characterize and
prioritize risk based on probability,
impact and preparedness
Risk insight and foresight: use business-specific scenarios,
stress tests, and early indicators to understand risks and
opportunities (potentially also for key customers and peers)
Risk models: build simple model as support tool
for business decisions
Risk reporting: focus on key risks
and provide clarity on these
to allow actionable measures
MACHINE LEARNING TOOLS CAN MAKE
IMPROVEMENTS IN THESE NINE AREAS
WITH TODAY’S “NEW” TECHNOLOGY
1
2
3
4
5
6
7
8
9
Chart 518
: Areas where new
technology can play

14
Chart 6 Risk Identiﬁcation Techniques (Navigant 2012)19
19. “Hope is not an effective risk mitigation technique” - Navigant (March 2012) - Brian C. Fox, James G. Zack Jr.
Technique Deﬁnition Example method type
Individual expert assessment The use of a single expert to identify
all of the potential risks on the project
through analysis and past experience
A senior project staff member
brainstorms all possible risks during
the estimating process and provides
information to the estimators
Tacit Knowledge
Multidisciplinary group assessment The use of a panel of experts to jointly
identify all potenial risks on a project
by relying on their individual expertise
Senior staff from the finance, legal,
operations departments as well as
outside consultants are brought
together to discuss and determine
potential risks for the project
Tacit Knowledge
Structured/expert interviews Formal interviews of subject matter
experts, combined with collected data,
are used to identify risks
An estimator interviews internal or
external experts and collect their input
in order to have a list of risks that may
potentially impact the project
Tacit Knowledge
Delphi Technique A highly sturctured, iterative expert interviewing method in which multiple
experts provide answers, receive feedback based on the collective response of
all the experts, and then are given a chance to revise their answers based on the
feedback. The process iterates until a preset stopping point
Tacit Knowledge
Checklists Use of pre-established list of items,
from which risk pertinent to the
current project would be selected
The operations department maintains
a 10 page list of potential risks that can
occur on any infrastructure project,
from which a subset of potential risks
for the current project are selected
Explicit knowledge
Risk records Past project records are reviewed to
determine what risks occurred on
similar projects in the past
Change order logs for prior projects
are reviewed for unanticipated
conditions
Explicit knowledge
Prompt lists Instead of providing a checklist of
risks, a list of questions is used to
focus attention on potential areas of
risk
Senior management maintains a
list of questions to be asked at the
commencement of each major phase
of the project
Explicit knowledge
Free and structured brainstorming A semi-structured thinking session in which all potential risk sources are
considered and subsequently assessed to determine if such a risk has the
potential to occur on the project
Analysis
Assumption analysis The identification and documentation of assumptions made during the
planning process, and the identification of potential risks that would result from
the failure of any of these assumptions
Analysis
Pondering ‘what could go wrong’
analysis
An unstructured thinking session
with one or more parties, where all
involved attempts to think of possible
risks
An informal meeting of project
managers is called in which the
preliminary budget is drawn up and
parties perform a line by line review to
determine potential cost risks
Analysis
Diagramming techniques The use of graphical techniques, such as fishbone (cause and effect) diagrams
to structure thinking about potential risks on a project, which is typically
performed during the project planning phase
Analysis
Synectics A technique similar to brainstorming, but more involved. The focus of the
method is to use comparisons, classifications, metaphors and analogies to
correlate seemingly unrelated ideas in order to bring to the forefront that would
otherwise not be incorporated
Analysis

15
Conclusion:
What does this mean?
Risk management and dispute resolution professionals are
realizing that advanced technologies are not just for game
shows, automobile navigation and shopping. They now have
affordable cutting edge technology choices beyond email to help
them capitalize on the deluge of data that harbors both risk and
opportunity. This means that executives, lawyers, arbitrators,
engineers, quantity surveyors, contract administrators and
risk managers, can work together to permanently store their
collective knowledge into “analytic brains” and create a
learning platform for a shared, collective, machine-assisted
intelligence. Arbitrators and arbitral participants can uniquely
benefit from “intelligent agents”, deployed during a contract’s
execution or after disputes have emerged, because their roles
depend on accurate “facts reconstruction” from warehouses of
data in some cases decades in the making. With these tools,
arbitral players will experience dramatic expansion of time
and money to more efficiently discern what is fair, indeed, to
do what they do best. This vision will not lessen the demand
for or importance of this community’s work. Instead, the
most sophisticated knowledge professionals in the world will
be better able to leverage their intellect instead of wasting it
with keyword search and storing documents in folders. As the
former CEO of Hewlett-Packard said:
“If HP knew what HP knows we’d be three times
more profitable.”
More simply, the business benefit of the future state is not really
about costs savings, or making the lives of the legal community
‘easier’; it’s about managing risk more effectively, removing
unnecessary big data driven legal friction from our economic
and legal systems. The legal community will ultimately benefit
from businesses’ improved financial performance. Investment
in these technologies will enable all players to build and take on
more and more audacious projects, to make the most efficient
use of their capital, and generally “do what they do best” to the
benefit of their companies and the world.

16
About the author:
Shannon Copeland is an advisor to Secretariat International,
an international construction risk management consultancy.
Shannon began his career as an offshore oilfield engineer
with Chevron and for the last eight years has led consulting
engagements aimed at using technology to reduce legal risk and
dispute costs. Shannon lives in Atlanta with his wife, Anna and
three children.

Smart Contracts AI Article

More Related Content

What's hot (20)

Similar to Smart Contracts AI Article (20)

Smart Contracts AI Article