Coming soon: full-blown AI in Weagree - Weagree

Coming soon: full-blown AI in Weagree

We have been working on AI since mid-2019. And then one year ago, ChatGPT made all then-existing AI-solutions almost worthless. The ‘trained models’ or mini LLM’s (large language models) of existing AI-tooling for the analysis of an NDA or other commercial contract, such as Project Atticus, were at once useless

Completely worthless or useless?

No. We developed our AI-tooling as a completely integrated part of contract automation and CLM. And the contract automation-Q&A (for contract creation) is entirely reusable in Weagree CLM (for managing signed contracts). The effect of this deep integration has two impactful advantages:

  1. No implementation is needed (Weagree AI will be plug-and-play prompted)
  2. Only those authorised to use it, get access (and access is in the natural flow: fully intuitive).

This second advantage is thanks to Weagree being enterprise grade, with advanced role-based user management. The fascinating part is in the first advantage:

AI automated NDA review automated analysis of an NDA

No implementation (plug-and-play prompts)

The deep integration of AI with Weagree implies an exceptional time-saving on the implementation work needed to start using AI effectively. The absence of implementation work accelerates the adoption of AI and accelerates every contract stage:

  • Automated risk assessment of first-draft contracts received from the counterparty
  • Automated contract review for red flags, and weighing of proposed terms against contracting policies
  • Semi-automated marking-up and red-lining with clause library clauses
  • Automated data extraction upon registering a signed contract into the CLM

It sounds too good to be true. How is this possible?

AI ‘prompt engineering’

The crux is largely in prompt engineering. ‘Prompt engineering’ is a somewhat hyped term for a rising legal discipline: the art and skills of formulating the right instruction to ChatGPT( or any other LLM – large language model). And the nice thing is: in Weagree, this is already implemented.

Alongside the contract you upload into Weagree for AI-review, we send the Q&A-elements used in the CLM into ChatGPT as well. And in a few seconds, ChatGPT returns three things:

  1. For each Q&A-element, its value in the uploaded contract.
  2. An explanation of why ChatGPT derives at this value of the relevant Q&A-element.
  3. The contract wording on which ChatGPT bases its analysis.

To make end-users feel comfortable, Weagree will highlight this returned contract wording in the contract text: in our famous editable under-water-screen. (See the pictures in this email.) All contract metadata are collected in Weagree’s automated risk assessment and Weagree CLM, and ChatGPT’s explanation shows up alongside the values/metadata.

AI based automated contractt review GPA automated analysis of an NDA

Integrated end-to-end contracting

We found out that ChatGPT is already exceptionally accurate if the prompt is limited to your Q&A-elements implemented for contract automation or CLM. The fact that Weagree is an integral end-to-end contract solution, capturing the entire contract lifecycle, pays off in adopting AI.

Questionnaire elements are not only used for creating a first-draft contract. In Weagree, you would reuse many of these Q&A-elements also for managing signed contracts in Weagree’s CLM. (This reuse of Q&A-elements explains why it takes only one mouse click to implement CLM, the post-signing managing of contracts.) And of course, as there is no 100% match between the Q&A for contract creation and managing signed contracts in a CLM, you can always enrich each CLM contract sheet by adding your own CLM metadata fields.

All this means that if you automated your contract templates in Weagree (and more so, if you did that on both the buy-side and the sell-side, you have everything ready to start using AI-based automated contract review, AI-based automated risk assessment, and AI-driven contract data extraction).

Our AI roadmap: AI-automated redlining

We are now building the features to (automatically) create a second-draft contract mark-up: the redlining of the received contract, with your own clause library clauses replacing unacceptable proposals. We are also optimising the user-experience so as to accelerate the contract flow.

Furthermore, we will introduce smart ways to optimise prompt engineering: this art or skill may be a hyped term, it will become a key discipline in contracting.

The game is on: for you!

And yes, while this is a spectacular milestone that puts Weagree on the stage, we will surely look for other AI-features. We reserved time for it on our short-term roadmap. The game is on and we are passionate to play an impactful role.

Of course, our customers can use their own instance of ChatGPT or any other competitive LLM (AI-trained large language model), and also this would be plug and play.

More soon! We are release-testing our AI-components, but the first results on automated reviews of general purchasing agreements, NDAs, SPA’s, LMA credit facility agreements are impressive.

If you like to join us in our fun now, let us know:

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: