Litera has launched a GenAI-driven KM tool that allows firms to tap prior work to find ‘relevant precedents, deal points, and what is market’ to boost a client’s negotiation position and support transaction planning. Artificial Lawyer spoke to Head of Product, Adam Ryan, to hear more about Foundation Dragon.
The company has stated that this LLM-powered capability, which taps into ‘all matter, deal and negotiating data’ from previous transactions – which have been placed in its Foundation KM system – significantly reduces the amount of time lawyers and associated KM teams need for this important work.
In this case, Litera reckons Dragon will save firms around eight hours of work on each prior deal analysis project, now doing it in minutes, which as Ryan explained, frees up associates and knowledge lawyers/PSLs to do more valuable work.
It’s not the first LLM capability at Litera, where Kira already provides ‘smart summaries’, and it’s clearly the wider direction of travel for the company.
In this case, they’re using OpenAI’s GPT4-Turbo to pull out the relevant data, but only after it’s already been ‘chunked’ i.e. made into smaller textual pieces, by Litera’s own software. Plus, the data has already been structured inside Foundation – and that’s a key point here.
Ryan said the result for Dragon is very high accuracy. He recounted a story about one firm that piloted the tech which quickly found that it could work better than some of its associates, who had missed key elements in past deal documents, but which Dragon did not.
It can also work in tandem with other Litera capabilities such as Clocktimizer, which helps to extract and analyse work costs for matters, allowing a law firm to build a very comprehensive picture of what any new matter might look like in terms of price and total work inputs.
Customers will have a usage level fee structure. This is because GPT4-Turbo will be in play as well as Litera’s own software. The fee system will have different tiers depending on how many tokens you want to use. Understandably, Litera didn’t want to get into specific prices, but Ryan noted that the cost to a firm of an associate spending hours and hours of time on analysing so much data and trying to draw structured insights versus the cost of using Dragon was very different.
In short, where such work is unlikely to be billable, as it may even be before a full mandate is given, then there’s a clear economic, as well as time benefit. While if the matter is on a fixed fee, such research work can be folded into that – hence the sooner you free up the associates the better.
Ryan also noted that this can be used to help with pitches and RFPs, where often a law firm’s BD team are brought in to help develop the message that the lawyers have expertise with regard to specific aspects of a potential deal. Dragon, with Foundation’s help, ensures that firms don’t forget they have actually got relevant deal experience in certain areas that clients want.
All well and good. But, what about data security? Ryan explained that this is central to the Foundation system which is designed to recognise and accommodate client confidentiality, even to the point of masking client names and prior deal purchase prices.
Insights Engine
So, what does it all mean? As Ryan noted, this is an ‘insights engine’. And while Foundation was doing plenty of good by capturing transaction information, tapping generative AI takes it to another level – primarily because the LLM of GPT4-Turbo can work in such a contextual way.
For example, Ryan noted how using Dragon you can discern trends in M&A deals – at a very granular and clause level – which would require a lot of hard work by a team of associates to do manually. In fact, it’s so useful for being able to capture a lot of insights it can even be used by a law firm’s marketing team to support thought leadership pieces for clients about market trends.
In fact, the longer Artificial Lawyer spoke to Ryan the more it became clear that this is A) just the beginning of the LLM-driven insights engine approach at Litera, and B) there are a myriad other ways it can connect to additional data sources and leverage that information for new insights.
In short, it’s taking a more all-encompassing approach to KM and turning it into actionable intelligence, which can then be used to help redesign the entire work process inside a practice group or team – which has to be the way to go (see earlier piece today: ‘GenAI Has Split the Legal Information Atom – Now What?’)
Then think of what else Litera does across its raft of products, and now think of every KM source really being connected to Dragon, feeding insights across the platform, with an LLM drilling away for you whenever you want some more granular intelligence, or to discern a pattern or trend. Interesting times for legal knowledge management, that is for sure, and it’s thanks to GenAI.