Which transactional lawyer doesn’t want to be able to tap their firm’s prior deals for knowledge that can help them today? Naturally, they all do. The goal of US-based legal tech startup Centari is to make this happen and Artificial Lawyer caught up with Kevin Walker, CEO and co-founder, to find out some more.
First, let’s look at what Centari does. Using genAI to tap and organise prior deal data, it seeks to help lawyers:
– ‘Unlock more deal flow – Don’t just tell clients you’re the best firm for the job. Show them. Enrich marketing materials and win RFPs with experience data that illustrates your unique expertise.
– Command the negotiation chessboard – Visualize market trends, benchmark your opponent’s previous positions, and achieve optimal terms for your clients with powerful competitive intelligence.
– Never dig for precedent again – Experience next-generation precedent search, powered by structured data. Centari enhances documents with rich metadata to pinpoint the exact deal you’re thinking of.’
So, how did this all get started? Walker explains that he worked at leading US law firm Paul Hastings, but left after about a year, then worked at several businesses, including as the GC for a startup.
‘I always had the itch to do something entrepreneurial and I could see that so many things could be improved [with lawyers’ deal intelligence], and that so much could be automated,’ he says.
As generative AI arrived, Walker realised what was possible. He soon met CTO and co-founder Bryan Gilbert Davis, and in March 2023 Centari was launched.
‘After ChatGPT it was clear that it was now or never,’ he adds. ‘I saw the opportunity to build something that would help lawyers with their data strategy, that would maximise the use of their own unstructured data.’
He also notes that as a GC he ‘had to purchase legal tech and I was frustrated that the companies didn’t empathise in terms of how hard it is to be a transactional lawyer’. Centari is the result of all those experiences as well.
The Product
Walker notes that companies such as Henchman (now part of LexisNexis) tap a DMS to find clustered clauses that connect to something a lawyer is drafting, but they want to be more comprehensive than that.
As part of the above features, the aim is to provide lawyers with ‘what’s market’ across the transactional field and offer key facts about multiple aspects of certain deal types, such as what is the current length of time for the earn out provision in a certain industry sector?
‘Knowledge is power, and firms have that [data], but it’s unstructured. Once it’s structured and clean, it can be used for all sorts of things,’ Walker states.
In short, the philosophy is that once you’ve got the right data, extracted it and made sense of it, then it can be used in multiple ’downstream use cases’. It’s just a question of how you want to leverage it.
And that makes a lot of sense. It’s an approach that Litera’s Foundation Dragon product is also taking.
In terms of the tech, they have developed a broad multi-model engine that taps different genAI systems, from Gemini, Claude, and the GPTs, to some open source ones as well.
‘We are very flexible and some customers want everything to run on their own servers, so we can use open source there,’ he adds.
He also mentions that they are ‘benchmarking accuracy for each one’ and that they are keen to be part of a move toward transparency across the sector when it comes to genAI tool performance.
Where Next?
They have done a pre-Seed $1.7m raise, with Jack Altman – the brother of Sam Altman – taking part, among others. They are also seeking further funding in order to grow.
However, Walker is keen to stress that raising cash is not an end in itself.
‘There are a lot of legal tech companies raising a lot at the moment, but we aim to build a product around the use cases, it’s not just about scaling up the company,’ he adds.
In terms of the big picture, Walker says the goal is to ‘build a full stack transactional law platform’, which can compete with multiple point solutions at once. In short, there is ‘an opportunity to verticalize the whole stack’.
Beyond that, there is the emerging field of agentic tools, where perhaps an AI agent can negotiate a contract for lawyers – not totally on its own, but certainly handle a solid part of it. They could tap structured negotiation positions and past data to then mark up the proposed contract, with the agents driving this forwards to an end point.
‘The agent could even understand how the lawyer works [based on their past deal work],’ Walker concludes, and envisages a lot of additional developments in the field of legal genAI to come.
Exciting times and good to see new companies such as Centari seeking to push the envelope.