Software-as-a-service companies — the likes of Salesforce, Shopify and Zoom — are stalwarts of the modern business economy.
A London-based legal services startup, Avantia, wants to become a service-as-a-software provider for companies seeking a cost-efficient law firm. Avantia aims to merge the typical offerings of a traditional corporate-focused law firm with the emerging generation of legal technology software, specifically the tools built upon generative artificial intelligence.
Avantia’s business model is predicated on the notion that a law firm established on a foundation of using AI tools to automate lower-value legal tasks can yield sizable cost efficiencies that will attract legal clients to its fixed-price menu of services. For one thing, it has dispensed with the billable hour.
Avantia, which has around 100 employees and 90 clients, primarily in the asset management industry, is expanding into the U.S. with a New York office opening early next year. Avantia provides a typical range of routine legal and compliance services, such as drafting a contract or nondisclosure agreement, locating a legal precedent or reviewing a contract, with about 60% of revenue from the U.S.
“There’s a very significant opportunity to combine [legal] services with a very tech-enabled offering,” says James Sutton, Avantia’s chief executive. “So not be a SaaS business by itself, not be a services business by itself — be the thing that sits in the middle of that.”
Sutton, the former general counsel of asset manager Och-Ziff, now known as Sculptor Capital Management, founded the company in 2019, partnering with Oxford University AI researchers. Legal Dive spoke with Sutton about how AI is changing legal services, why GCs struggle to reap AI cost efficiencies, and why speed often outweighs price for many clients.
Editor’s note: This interview has been edited for clarity and brevity.
LEGAL DIVE: Many in-house GCs view legal AI software as a budget salve, automation to drive sizable efficiencies in their departments relatively quickly. Is that too simplistic?
JAMES SUTTON: The feedback I get from general counsel is two-fold. No. 1, it’s almost never doing what it was sold to do on the demo. And No. 2, you have to work a lot harder than you realize to get it to add value. So the AI companies are almost pushing a lot of that adoption stuff over to the GC, and saying, now go and figure out how to use this technology. That's part of the reason why we think our business model is interesting, because we in-house all that stuff. We figure all that stuff out for you. I suspect, based on conversations we’ve had, that a lot of the efficiency gaining is over reasonably low-level stuff. Like proofreading or checking the definitions in a contract before it goes out, doing a piece of legal research, or checking the content of a piece of legal research or tidying up the language. I think that’s probably where you’re seeing most of the biggest wins. I don't think you’re seeing significant cost efficiency and savings yet in the stuff that really is where in-house legal teams’ budget goes.
What is Avantia’s general pitch and the cost savings you tout?
Our clients care more about speed than anything else, so we’re saving them time to get the document done. For our toolkit at the moment, at the top end on things like NDAs, we are close to getting to around a 48% margin boost on the NDA contract. So it gives you a sense of the scale. There’s about 50% efficiency gain by using the technology.
Over time do you think we’ll see the large corporate firms stop doing NDAs and other lower-level work?
If you take the premise that AI will start to add significant efficiencies, then we should begin to take more of the routine layers that traditional law firms do themselves. And that pushes that high-value partner model up into strategic advice. It pushes them more into that complex legal work layer, versus trying to fight with businesses like ours at the routine layer. We're beginning to see that. We now have some work streams where our client is taking work away from one of the Top 50 law firms that work for them and bringing it over to us.
How do Avantia lawyers use AI? Is everyone trained on the tools or is there a separate, perhaps junior, staff that handles the AI aspect?
We’re very anti that model. Our technology sits in [Microsoft] Outlook. So we want our technology to exist where our people are. It sits in Outlook as a plug-in, and it sits in Word, and the two talk to each other. So if you do something in Outlook, the Word version of it knows what’s been done. And the technology, you interact with it via natural language. You talk to it. It's a bit like a chatbot. You interact with it by asking it to do things for you. Sometimes the workflows are AI driven: Go mark up this NDA. Sometimes the workflows are just engineering, and the intake system has to decide what workflow to run. For example, a popular one is, “Go run this redline of these two contracts for me.” There's no AI in that. The only AI is in the question. And then it’s recognizing a workflow that it knows how to run, which is just an engineering workflow running a redline, but it saved time. We are very against having our senior lawyers just have paralegals go off and run [the work]. A big efficiency gain we found was upskilling juniors. If you can take your paralegal cost unit, which is a fraction, something like 25%, if not less, of your lawyer cost unit, and you can upskill that paralegal to do more of the stuff that the senior lawyer would do through the agent toolkit, we actually find that that's one of the biggest efficiency gains that you get at the moment.
Do customers want to know the nitty gritty of where any efficiencies are coming from?
It’s a mix. It’s not part of our normal demo, but we will show them exactly what it does and how we use it. Customers just want you to solve the problem. That’s number one. They don’t care how you’re doing it. The one thing that they do like is that you are investing in technology and technology enablement. It allows our customers to say, “Hey, we are using AI. We are figuring this out, but we're not having to go through this whole adoption thing.” We're solving the adoption, figuring-it-out issues to some extent. But the reality is, they just care about: Can you solve this problem for me and how much does it cost and how fast can you do it?
What keeps you awake at night?
You’ve got ChatGPT 5 landing in December, and the future rounds of large language models, do they sort of catch up with what we’ve been able to do on a much more specific basis with contracts? It very quickly upskills a law firm that has not invested any money in technology at all to level play with us straightaway. That’s like the big existential question that we have to be comfortable with. We’ve got lots of reasons why we don’t think that happens, and we’ll continue to stay ahead of that wave.
