Having previously worked as a corporate M&A lawyer at Slaughter & May, and as general counsel for global asset manager Och-Ziff (now Sculptor) Capital Management, CEO James Sutton founded Avantia Law in 2019 with a vision of providing “accelerated legal solutions” for its clients in the asset management industry — by betting big on AI.
This month, the venture capital-backed ABS announced the launch of its purpose-built, proprietary AI agent Ava. Developed in-house by its team of AI experts, according to Avantia’s press release Ava “streamlines legal work, takes care of routine legal processes, manages lawyer workflow, and delivers faster, more efficient outcomes for clients”.
How would you describe Avantia’s business model?
I’d call it tech-enabled legal services, sitting in the space between law firm, ALSP and legal tech provider. We’ve based our business on the thesis that there’s an interesting opportunity right now to combine legal services and technology in a way that others aren’t doing.
Our business specialises in solving a small number of problems within legal services — contracts, compliance and LP transfers — for our clients. These tend to be high-volume tasks that are reasonably replicable. We’re investing heavily in how to solve these very efficiently using AI, and when we feel we’ve got to a good point there, we’ll look to grow and build out the range of legal services we offer.
How are you leveraging AI at Avantia?
We’ve built Ava, an internal AI agent that integrates with Microsoft Office via its API. Whatever happens in our organisation, whether it’s a lawyer corresponding with a colleague or a piece of client communication, our AI sees it and downloads all that data. We’ve used an open-source large language model (LLM) that acts as the AI agent’s interface, and our people ‘communicate’ with it. When instructed to run a job, Ava will route the task to other, smaller language models we’ve trained on our own datasets.
For example, we’ve got a unique dataset of 55,000 NDAs we’ve used to train a highly effective, proprietary model that our lawyers can use to ask Ava to markup an NDA. Similarly, it can run tasks on a predictive basis, such as drafting a reply to an email that comes into somebody’s inbox.
What impact do you think agentic AI will have on software use in the legal sector?
Software-as-a-service (SaaS) has limitations. These platforms are inherently built as generic solutions to common problems. What we now have is a proliferation of SaaS tools on your desktop, each with its own user interface. You can make them talk to each other, but they weren’t really designed to do that.
That’s changing with AI — we can now do away with these interfaces or artificial portals, and communicate with computers in our language.
For example, we bought a contract lifecycle management (CLM) SaaS tool which we’ve used in the first four years since we were founded. Now we have our AI tool, our lawyers don’t use the CLM anymore, because they can do all the same things — contract generation, storing workflows — simply by communicating with the AI in Outlook. The CLM is still being used as a data storage repository, but soon we won’t need it at all. It’s being completely replaced by AI.
How do clients view the involvement of AI in their legal work?
We’ve certainly moved well beyond that world of using open tools, like ChatGPT, that train on your data. All our AI models and data sit completely within our own infrastructure, and we have very high standards of enterprise security around their use.
It’s no different to how you’d feel about inputting your data into third-party SaaS tools. It’s not that clients don’t have questions about it, but the general understanding at this point is that you can and should be using AI, while keeping your data safe.
How do you measure and price the value of the legal services you provide?
We have a pay-per-use model where we typically charge a fixed fee price. We don’t like the hourly-rate model because it creates all the wrong incentives for law firms, and we don’t think the subscription model is quite fit for purpose for legal services.
Fixed-fee pricing gives clients certainty of spend, and it incentivises us to get the work done as cost-effectively and as fast as we can — which makes us focus heavily on the technology investment to innovate on that front.
LPM firm facts: Avantia Law
Founded: 2019
Business structure: ABS
Headcount: 100+
Revenue: undisclosed
Office: London
How do you view the competition from large law firms, many of which are investing significantly in AI?
There will obviously be winners and losers as we all adapt to this disruptive technology. The opportunity I see for us is to innovate faster than traditional law firms can due to the nature of their ownership structures.
We’re slightly sceptical about the investment story and whether it’s really going at pace, given that the partnership model makes it extremely hard for traditional firms to allocate capital today for the benefit of the business tomorrow.
We’re also seeing many large firms buying third-party AI tools. Our belief is that because our AI system is so deeply embedded with our data, we are already significantly ahead of what these generic AI tools are capable of. They will obviously solve a much broader range of problems; that’s their appeal. But the combination of law firm plus generic AI will not be a match.
Even among the bigger ASLPs, their businesses are so complicated, with so many different services for different clients, that trying to bring order to that quickly is a mammoth task.
We’ve got the advantage of being a reasonably new business, where we’ve built and designed everything as we want it — solving a small number of legal problems very well, for a client base we know thoroughly.
As simpler types of legal work become more automated, isn’t there a risk of general counsel shifting more of this work in-house?
It is something we debate internally, and I don’t think anyone has the answer to this yet. The bet we’re making is that it’s the opposite — that we become so efficient at doing this work that we can take a whole business process from the in-house team, because it’s the most cost-effective solution for them.
Another advantage we have is access to much more data across a range of clients. We’ve got a huge database of engagement letters, for example, and can deploy that knowledge with our technology to give clients a solution that’s more sophisticated than what they could produce internally.
And, of course, there’s our secret sauce: how deeply Ava is embedded in our day-to-day work, and how effectively our people work with it to drive our efficiency. Our AI-enabled lawyers can work smarter, faster, and with greater impact.
What impact do you anticipate AI having on the legal profession in the long run? How do you see the relationship between the technology and lawyer evolving?
There’s an ongoing debate and concern that we’ll lose the next generation of lawyers as AI replaces the junior lawyer role. Our experience has been the exact opposite.
Because natural language interfaces allow people to interact with computers and data in an intuitive way, we’ll no longer need to train people on how to look up relevant information; it’s almost plug-and-play. By giving paralegals or junior lawyers access to the information they need when they need it, we’ve noticed their output improve more rapidly, to the equivalent of what you’d expect from a two- or three-year PQE lawyer. For training lawyers, AI is one of the biggest efficiency gains we’ve seen.