AI-automated review of the holy legal grail (LMA) - Weagree

AI-automated review of the holy legal grail (LMA)

Over the past weeks, we smoothened out all wrinkles in Weagree’s AI capabilities. Now let AI boom!

An LMA facility agreement is not the most complex type of contract automation, we know, but large (financing) transactions speak most to a transaction lawyer’s imagination. To extract all LMA-parameters from a first draft prepared by the other party, alongside an executive summary within a few minutes. (Wow!) See the screenshot on how intuitive Weagree AI is:

AI review LMA highlight fields 640 01 AI-automated review

So, here is your dream-came-true:

  1. Upload. Get the first-draft LMA agreement into your Weagree portal and initiate AI-review. There are several ways to upload or import, but forwarding your email is an appealing one: Weagree handles the forwarded email-with-LMA-attached mostly automatically.
  2. Convert format. Weagree AI automatically converts the Word-file into a text-format that works both for GPT 4-Turbo and in Weagree’s in-app contract review-and-negotiation tooling. (Reverting it back into Word is easy-peasy as that is our raison-d’être.)
  3. Identify parties. Weagree AI also automatically identifies the contracting entities: for each, the party details are automatically retrieved from Weagree’s in-app CRM, an external CRM or added if new.
  4. CLM review-fields. For reviewing an LMA, there are easily 300 relevant contract parameters: as we are reusing the LMA-creation questionnaire to AI-review, setting up the AI-prompts takes hardly any time (6-8 hours). We selected the parameters of an LMA Term Sheet for creating the right ‘executive summary’ transaction data. We rearranged them slightly in our CLM, as you can see in the screenshot.
  5. AI-driven risk assessment. For each relevant LMA parameter, a risk rating can be defined, as well as the nature of the underlying risk. This is applied automatically, and enables end-users to identify key negotiation aspects with the blink of an eye.
  6. Spectacular manual check. As GPT 4-Turbo returns both the contract clause text and the explanation as to why GPT believes the returned value is.
  7. Click around. Once the AI-review is returned from the AI model, the end-user can simply click around data, contract text, risk ratings, and back.

Here is another screenshot of how accessible (and manageable) an LMA transaction is:

AI review LMA transaction view 640 AI-automated review

Do I need to explain how thrilled we are?

Our roadmap to make Weagree AI perfect

We were thrilled by how much is already possible and how little implementation efforts are needed to start using AI-automated contract review. Our latest testing and benchmarking tell us that most of what we need to develop is limited to application-infrastructure components (webpages with the right features).

Therefore, here is our roadmap for Q1 2024:

  1. In-app contract negotiation tooling upgrade: options to modify the contract in track-changes within Weagree (incl. distinguishing for Own party vs. Other party), and options to ‘comment’ (negotiate) the contract text (distinguishing internally vs. visibility to the other party). Expected to be ready end of February.
  2. CLM features. We will strengthen the CLM (post-signing) search functionality with full text search (incl. automated OCR on uploaded PDF’s), fuzzy text search (search contracts for a fuzzy word), tagging of contracts and files (incl. related tag-search options). Expected for mid February.
  3. AI features. To be further elaborated and expanded: structured AI-summaries of the reviewed contract, AI-summarise clauses (with the option to comment on it and discuss it with colleagues), AI-driven modification of contract provisions (using the above in-app contract negotiation tooling), options to approve deviations from contract requirements. Immediately after completing the in-app negotiation tooling (March 2024).
  4. Clause library insertions in this in-app negotiation tooling: insert own contract language in the text provided by the other party (and create the track-changes). After completing the in-app negotiation tooling (March 2024).

As you can see in the screenshot: the user-interface is really intuitive:

AI review LMA highlight fields 640 02 AI-automated review

Weagree AI – our use cases

What applies to the large LMA financing agreements is more easily achieved for other types of transaction or documentation:

  • M&A-related SPA’s, APA’s, SHA’s or JVA’s.
  • FIDIC suites of construction-related contracts.
  • Legacy CLM contracts (large-volume for automated registration in Weagree CLM).
  • Due diligence (automated review of large sets of mixed-type documents).
  • Migrating legal entity management files (large sets of deeds, documents, certificates and files related to corporate housekeeping).

Neither of the above use cases would take us much development time. Still, we will only pick them up (and prioritise them) together with customers who are eager to collaborate with us in realising it.

Here is a peek of LMA-related CLM data fields (at the right-hand side are the explanation how GPT came to its review, and if you click on it, it will jump to the contract clause (as in the above screenshot):

AI review LMA CLM view 640 AI-automated review

No implement efforts – how come?

Implementing a comprehensive AI-review of LMA agreements takes merely 6 to 8 hours to implement. All follow-up then still needed is: finetuning and optimising the AI-prompts (this may take another two or three days, because with LMA transactions, there are easily 300 AI-prompts). Configuring the risk rating may take another 3 to 5 hours.

It is so easy and little-effort-consuming thanks to having your questionnaires in place! Here is how an authorised user can adjust the risk rating returned by the GPT 4-Turbo:

AI contract risk assessment adjustment AI-automated review

Thrilled as well? Want to reflect on AI-automated contract review, AI-contract risk assessment, or about simple AI-driven registering contracts in your Weagree CLM?

Terms of Use

I hereby accept (or reconfirm my acceptance of) Weagree’ Terms of use, in which:

Terms of Use

I hereby accept (or reconfirm my acceptance of) Weagree’ Terms of use, in which: