The Most Overlooked Factor to Generative AI’s Potential

By Alex Smith.

When we witness the revolutionary results that generative AI can serve, it’s easy to forget that the bulk of that “magic” is merely an extension of the quality and consistency of data, people, and processes supporting the ongoing curation of that data.

Given data’s foundational role in AI, it’s important for organizations to ensure they take a holistic approach to data collection and curation if they hope to achieve the best results.

Capturing an overlooked knowledge asset

Many legal organizations have perfected building and maintaining a knowledge library – home to templates, playbooks, and other knowledge assets that represent the best work and best thinking – to ground their AI. But they often overlook certain types of knowledge. Take “closing bibles” or closing sets, for instance.

The closing set is more than just a final record of an M&A transaction or some other high-stakes deal – it is also a key knowledge asset waiting to be tapped at the final stage of the deal lifecycle.

Throughout most of a matter, legal professionals will have been working on documents within a document management system (DMS), participating in negotiations with clients, and collaborating on different versions of the documents until they’re finally ready to close the deal.

Once all the signatures are in place, the closing set is created and eventually saved back into the DMS as an “object of record” final version. But too often, that’s the end of the road for that closing set, left to gather dust in the DMS.

Instead, organizations should have a process in place – either manual or technology-assisted – to push that closing set into the knowledge system. From there, the organization can run AI over the closing set to accomplish several things. For starters, AI classification can determine exactly which types of documents (e.g., purchase agreement, nondisclosure agreement, and so on) are in the closing set, which is no small task, given that closing sets can often contain 30-50 different documents.

Once AI has lent a hand by classifying the different document types within the closing set, it can take a deeper dive and see what type of interesting or creative drafting it includes. How was the deal structured? Are there any interesting clauses in there that could guide legal professionals who are facing a similar deal in the future? You could even ask a gen AI model to review a handful of different closing sets for you and advise which one would be most relevant or most useful for an upcoming deal.

Your infrastructure impacts your data

By now, it should be clear that success with AI depends on your data infrastructure– and that data is only as good as the people, processes, and technology that the organization has in place to capture and leverage it.

For example, the last thing an associate wants to do at the end of a busy deal is spend half a day putting together the closing set – particularly if the senior partner is already all over them about another urgent deal. There needs to be a process in place that streamlines and automates not just the creation of the closing set, but saving it back into the DMS and moving it into the knowledge library.

At the same time, AI can’t be viewed as something that happens in a vacuum. If materials like closing sets need to be taken out of the DMS in order to run AI on them, that introduces security and governance risks. Not only that, once data is moved from a system of record across different systems/platforms, the data quickly becomes disorganized. Firms can wind up with duplicate systems of record where it becomes unclear where associates should pull the latest version – and they lose time sorting out whether the version they’re looking at is the correct version or not.

The lesson? Your infrastructure impacts your data. Firms need to take a holistic approach to their people, processes, and technology to make sure they can get their data into the right places at the right time – which is the only way they’ll have a proper foundation for their AI to draw upon.

Application: create a formidable foundation

Ultimately, success with AI is less about the actual AI tool, and more about the data it is drawing upon. It’s the bedrock of AI’s brilliance, which is precisely why firms should leave no stone unturned in their quest for quality data.

Firms that can close the loop and make every ounce of collective intelligence – even under-utilized assets like closing sets – something they can leverage will be positioned to deliver the best outcomes with AI. The investments they make now in creating this foundation of quality data will only continue to benefit them as they move forward.

---

About the Author Alex Smith is Senior Director – Knowledge, Search, AI – at iManage. With significant experience in developing early-stage and innovative digital product concepts, he is responsible for managing the product roadmap for the company’s Search, Knowledge, and AI offerings. Leading major customer discovery projects at both market and specific product levels, Alex has helped to build user-centric capabilities and aid the adoption of new concepts and ways of working in all types of legal organisations. Before iManage, Alex held innovation-led positions at Reed Smith LLP and LexisNexis.

#AlexSmith #AI #data #lawfirms