What Gen AI will bring you - Weagree

What Gen AI will bring you

It is AI all over. Which generative AI (Gen AI) solutions will arrive first, and what’s next? You wonder where to start with AI regarding contract automation and CLM? Here are a few thoughts.

For Gen AI, the main contracting use cases will be:

  1. Contract and clause summarisation
  2. Contract and clause translation (in multinational settings)
  3. Contract creation (or clause library)
  4. Chatbots
  5. Precision redlining (automated mark-ups)

It is not us concluding this; it is last week’s CLM/Gen AI research by Gartner that anticipates this.

Maturity and risks

How each of these use cases will reach you will depend on its level of maturity and the perceived risks involved in adopting them (set off against the costs and efforts to implement and operationalise).

On the short term, ‘off-the-shelf AI solutions’ will likely be more successful as they are already very affordable. On the long run, more organisation-tailored AI solutions may gain foothold; but it still requires that AI training or machine learning are much more automated.

What does maturity mean?

Gen AI technology is likely to mature relatively quickly, thanks to an increased acceptance for AI-based solutions (and increasing pressure to embrace AI solutions) combined with the increasing quality or decreasing risks.

Successful change management in optimising legal operations has become a prerequisite. It means that stakeholders involved in contracting must have the right mindset: open to adopting technology. Transitioning to AI is easier if transaction management (the precontractual stage of creating, negotiating and signing a contract) or CLM (post-signing) had already been automated. The fresh jump from spreadsheet-contracts-management to AI-driven automated review, risk assessment or CLM is simply too big.

Most AI applications are still immature. This has two aspects: the quality of the AI model vs. how a CLM application (granularly) embeds AI.

AI review LMA highlight fields 640 02

Quality of the AI models

We expect that the quality of large AI models like ChatGPT are improving so quickly that they will overhaul tailored solutions (user-trained AI models). The legal world will eventually adopt end-to-end contracting solutions integrating with such proprietary versions of ChatGPT and the likes, rather than developing or customising such solutions with their own in-app trained AI model.

And yes, we will soon also see hybrid AI solutions: combining the best of an integrated ChatGPT and their own (partly automated) tailoring options. If AI training, machine learning or customisation is user-friendly and easy, these are the winners.

Quality of the AI-embedding applications

Maturing will require an optimised embedding of AI technology in the contract core processes. We will see a more differentiated application of AI use cases, more granularly embedded with contracting steps or contract management needs.

In other words, contract processes will be broken down to smaller steps and each such step accelerated by automation or AI. That is what Weagree has been doing for over 18 years.

Which AI use case first?

The first two use cases (contract or clause summarision and contract or clause translation) are relatively easy to adopt. If you have worked with ChatGPT or DeepL, you appreciate that ‘AI performance’ is already of a surprisingly great quality.

As the reliance on the AI output (a summary or a translation) is sensitive but not that crucial, risks associated with such AI applications are perceived as relatively low. Moreover, the benefits of these two use cases (summarising, data extraction and translation) are considerable, as it will elevate contract management practices to a significantly higher level (I explained this in an update of a few weeks ago). This combination of ‘low risk’ and ‘higher performance’ will drive adoption of these AI use cases.

The trick for these two use cases is to handle the AI-resulting summary or translation properly: making them accessible and reportable.

And no surprise, contract creation (incl. clause library) entails translating human questions into the right contract and the right provisions. In other words, the AI elevates a contract creation questionnaire to a higher level or the AI results in plugging in the right contract building blocks into a markup. Given Weagree’s strong foothold in contract automation, it is the area where Weagree has impressive potential.

Also precision red-lining (replacing a mere sentence by your own) is something Weagree will offer shortly. It is just a few small steps beyond what Weagree has been doing for 18+ years already.

Chatbots will prove to be the most challenging: the combination of required customisation, a probable lack of ready-made (context-specific) chatbot-answers (availability of full-fledged playbooks, contract know-how and written sector-specific knowledge) and the associated risks is likely to be costly.

Weagree and AI

Obviously, we are aligning our roadmap accordingly. We are inspired by a research paper of Gartner of last week. We have been experimenting with AI for over four years.

It means that the Weagree interface is already well-thought-through and the granular integration options are well-developed. Non-AI aspects, like AI-based contract risk analysis or automated import of Word-files require technology that we know ‘better’ than anyone else (as we are 18 years in that field). We are finetuning user-friendliness rather than overhauling our contract automation or 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: