Intapp this month announced that its agentic ‘AI coworker’ Celeste, previewed in February, is now generally available. The listed company positions Celeste as part of a new category of ‘firm AI’, focused on improving law firm performance rather than individual lawyer productivity.
Celeste sits in the business of law space that Intapp has operated in since its inception, helping firms with deal screening and conflicts clearance, to business development and lateral hiring.
Speaking with Legal IT Insider, co-founder and chief product officer Thad Jampol said: “Every desk, every lawyer is using some form of AI. It’s flooded the market. And I think it is having some beneficial effect that lawyers are drafting faster, they’re reviewing documents faster, they’re working on contracts faster, certainly more than they did a year ago or two years ago. But the question everyone’s asking is, is the firm growing any faster?
“For most of these organizations, the answer is not yet. And it’s really because AI built for the practice makes individuals faster, but it’s not converting into growth for the firm. So, the headline is that individual productivity is just not equal to firm productivity.”
Jampol says that almost every tool up to this point in the generative AI, agentic AI wave has been built for the individuals and not for the firm. He points out that there are three ‘categories’ of AI emerging: generalist assistant AI such as Claude or Copilot (“that we think every firm is going to have and is really good for everyday work”); then the practice of AI, such as Harvey and Legora (that “have had a lot of oxygen in the industry first wave”); and a third category that is firm AI for the business of law (“and no-one, almost no-one, is building that right now because it’s the hard one to do.”)
Celeste focuses on core workflows or value chains that Jampol describes as “the beating heart of the firm as an organization.”
He says: “In a really interesting way, those key workflows have not changed over the years. Firms have to go win clients. They have to expand their relationship with those clients. They pursue laterals. They have to go through a lot of very onerous, ethical, financial, strategic steps to onboard those clients, upholding their duties of loyalty and their ethical obligations. They have to staff the team. They have to then go deliver the work, but also get time out and ensure that it is invoiced and collected. And that end-to-end client matter lifecycle, we believe, still holds.
“Now, it may begin to be re-engineered as from a process standpoint to really take advantage of the agents and the new technology. But we’re really bringing, I think, a totally new set of capabilities to those same important workflows. A lot of the past technology generations, like if you think about when workflow and BPM came out and then like robotic process automation and RPA and there’s a lot of these technologies that certainly helped and Intapp used, but a lot of times they worked around the real substance of the problem. They were really about the routing of information and then the approvals and the reporting of that and the audit logs. Now with AI and agents, we can really begin to solve more of the day-to-day decision-making and judgment part of the challenges that I think are really going to take this to a totally new level.”
It will appeal to firms that Intapp’s pedigree is in Walls, and that Celeste enforces your guardrails.
Jampol says that the market reaction to Celeste has been extremely positive, observing that, while everyone is talking about AI in the practice of law, “everyone has the same thing. Everyone is using the same models. It’s kind of become table stakes. The firm isn’t seeing an improvement in financial performance or competitive advantage.”
The ability to take processes such as bringing on lateral hires and use AI for vetting, conflicts checks and managing the integration is, Jampol says, “a really big needle mover.”
Another early use case is on the intake and conflict side. Jampol says: “It’s the process that every firm has to do that stands between winning a client and then being able to start doing the work for them and billing for that time. And anything you can do to both shrink that as well as take off a lot of the onerous work that the partner themselves has to do instead of spending time with that client, they love it.”
Baker Hostetler is one of Intapp’s Celeste launch clients. “That is the use case that they really grabbed onto,” says Jampol. Hg is another named early adopter. Shirin Veeran, head of data, tech & AI strategy at Hg, says: “The intelligence is ours, not something generic applied to our firm. Celeste screens against our actual mandate, drawing on knowledge we have built across the firm over two decades. Having a clear, always-current view of deal progress and pipeline speed — without the usual reporting lift — is genuinely valuable to any deal team.”
In terms of whether this is a new category, Litera has long focused its AI developments on the business and practice of law and with its recent relaunch, undoubtedly focuses on the ‘firm AI’ category.
As margins come under pressure, however, it’s undeniable that there will be more focus, from different lenses, on firm-wide efficiency. Many law firms find that their business services resources are stretched, and anything that can truly enable them to augment overstretched teams is likely to prove attractive.
