As a global professional services business, Accenture houses more than a million contracts in its records system, with thousands being added every month. In the course of contracting with a client, Accenture attorneys sign numerous initial documents, master agreements, and possibly hundreds of change orders and statements of work. In their day-to-day work, they have a need to understand Accenture’s rights and obligations across contracts and precisely how they are documented. This need requires being able to find specific contracts and extract relevant information from them.
These seemingly simple tasks, however, were too often manual, tedious and costly to perform because the ability to do detailed searching in the current system was limited. Given this situation, Accenture Legal turned to a CIO Applied Intelligence team from Accenture’s global IT organization for a solution. CIO Applied Intelligence drives innovation and applies predictive models, AI and machine learning to products to bring greater insights to the business.
When tech meets human ingenuity
To develop a solution, Accenture Legal, CIO Applied Intelligence and the records system application team focused on addressing two needs: enabling the capability to search for a specific keyword and to search for specific clauses across contracts within Accenture’s contract records system.
The CIO Applied Intelligence team designed an intelligent and robust contract search tool as a solution consisting of two core components, each addressing the two defined search needs with new functionality. The solution would also serve as a contract search repository. They named the new solution ALICE, for Accenture Legal Intelligent Contract Exploration.
The two core components developed were:
The CIO Applied Intelligence team developed standard text searching of search terms consisting of one or several words. An authorized user enters a term in the search field of the UI. A call is made to a cloud-based content search service, and the results containing a list of all documents containing the search terms are returned from that call and populated in the UI. The user can download a specific document from the list of documents in one of two ways: in a Word format or in the original form, which is generally a scan of the original document in PDF form.
For the search for contract clauses, the team developed a clause extraction solution based off finding keywords to identify passages in a contract document that reference particular clauses, such as, for example, force majeure, limitation of liability or data privacy. Because these clauses may not be referenced directly or use the same words, related keywords are used to allow a way to identify when a clause is present and to identify the different clause types.
The keywords are identified through "word embedding," a natural language processing (NLP) method and part of an emerging field of AI called deep learning. The team used a word embedding model built on Accenture’s contracts to find keywords with the greatest similarity to the different clause types.
The result is a capability to search for any kind of text a user is looking for in a contract document and to search for clauses within a document. ALICE continues to be "trained" to become increasingly more precise and relevant in search results.
People and culture
Today, Accenture Legal’s global team of approximately 2,800 professionals is using ALICE. Users can more efficiently perform precise information searching and extraction, unleashing data that was previously not easily accessible.
Making the data easily accessible to Accenture Legal is delivering huge and immediate benefits. Chief among those are the ability to quickly search vast numbers of contracts for risky terms, which can be written in multiple ways, and the ability to quickly check all global contracts in the event of a flood or other unexpected local event and provide an indication of impact and risk to senior Accenture leadership. In the past, gathering this information was a time- and resource-intensive effort that occurred a few times a year.
ALICE is also helping legal professionals use the information they extract to answer strategic questions, perform further analysis and observe contract trends. An additional and unexpected benefit has been to other parts of Accenture beyond Legal that are now gaining insights from Accenture’s contracts that were previously unavailable to them, such as trending information on new technologies being sold.
"It’s about how do we make a difference when we’re trying to give insights to our business leaders around the types of risks that exist and how we can best manage and mitigate those risks?"
— Christina Demetriades, General Counsel – Europe, Accenture
A valuable difference
According to Christina Demetriades, Accenture General Counsel Europe, "We’ve invested in ALICE because at the heart of it, it’s about knowledge: how do we have clear insight and knowledge into what is in our portfolio of arrangements with our clients? It’s about how do we make a difference when we’re trying to give insights to our business leaders around the types of risks that exist and how we can best manage and mitigate those risks? You can only do any of that if you have knowledge. At the moments that matter, when we need to find out what’s in those contracts or what exists in our portfolio, ALICE allows us to do that in an instant."
The deployment of ALICE is an exciting step in Accenture Legal’s journey to applying artificial intelligence for the Legal function. Enabling a repository of contract data creates huge opportunities to create further business value. The long-term aim is to create greater insights across Accenture’s legal landscape to enhance how Accenture manages the delivery of a contract and better manage risk overall.
The business value the Accenture Legal Intelligent Contract Exploration delivers includes: