Artificial Intelligence in Contract Management (Part 2: Knowledge Representation)

In the first part of this series, I introduced the topic of artificial intelligence (AI) in contract management. Today, I’ll touch on AI itself and the first of its areas that will be highlighted in this series: knowledge representation. AI systems are computer systems that emulate (and even surpass) human intelligence needed to perform valued tasks. The Turing test for an AI expert system is whether it is indistinguishable from a human expert that is communicating via, say, a “chatbot” interface. An AI-based CLM system would therefore be able to embody the collective intelligence of contracting professionals (internal and/or outsourced) into computer systems used during ongoing contracting strategy, negotiations and monitoring.

This scenario may conjure a vision of contract chatbots (or “bots” for short) replacing an entire industry of expensive attorneys, and indeed the technology will be disruptive over the longer term (especially for paralegals). But for now, AI is really about building collective intelligence to better model and manage commercial knowledge, thereby freeing contracting professionals to be more effective and perform higher value activities for the organization.

AI Elements and Their Relevance to CLM

To understand AI’s impact on CLM, let’s evaluate the various areas of AI and how they affect the world of contracts.

Knowledge Representation

Intelligence is meaningless without knowledge, and vice versa. Albert Einstein as a baby (high intelligence, low knowledge) would be as bad at chess as Wikipedia (high knowledge, low intelligence). Expertise is built on knowledge that adequately models the richness of a certain domain, but high intelligence allows the knowledge to be more effectively and efficiently applied to solve problems.

Contracts capture some level of legal, industry and organizational knowledge in them, but they are dumb documents. A contract document “object” has metadata (attributes or properties) that describes the contract (e.g., supplier name/ID, legal entity ID, contract renewal date) but not the knowledge that’s hidden inside it. This is why I outlined the first repository-building step of the three-step CLM journey as being “build a high-level knowledge base about all your contracts in the form of a contract repository to gain high-level self awareness of your commercial health vis a vis your contract documents.” As a side note, don’t wait to use CLM authoring and signing functionality until after you build your repository in gory detail across the enterprise. Fast-track your initial repository and then start establishing baseline clause libraries and contract templates as soon as possible.

Knowledge is not represented by documents, however! This is why my second main step in a CLM journey is to “derive key intelligence from within your contract data… to decipher the legalese down to a granular contract clause level (including metadata).”  This means expressively modeling very granular and inter-related data within/across the contracts via clause libraries and clause-level objects that have metadata, which describe business-relevant characteristics of those clauses: objectives, rights, obligations/commitments, remedies, rules, caveats/constraints, references (e.g., to external laws/regulations), risks and so on.

Note that it is also important to have legal and commercial experts establish “rules” regarding how best to manage the contract clauses and the contracts themselves. This allows for the concept of “guided contracting,” which I’ll revisit later when I discuss knowledge reasoning.

Also note that these various characteristics can have their own related object models and classifications (e.g., contract risks that have associated risk types, probability/impact calculation methods, mitigations, external data sources). Finally, there’s a commercial knowledge model and compliance-related knowledge model that “pegs” these contracts and clauses to various internal enterprise processes and external data so that the contracts are being actively integrated to counterparty/regulatory information and then monitored.

First Voice

  1. Contract Management:

    Very interesting article. While some of my colleagues have conversed about the future of humans in this context, I wonder about the future of the need for money. If AI constructs essentially expand the resources available to us to get valuable work done, it would seem that the need for money (and profit) may become obsolete.

Discuss this:

Your email address will not be published. Required fields are marked *