David Baffa, Lauren Leipold, Maria Maras and Timothy Fraser discussed how AI is reshaping the relationship between in-house counsel and their external lawyers during a session at the ACC's annual conference Monday. (Courtesy of S.Dressel-Martin/ACC)
During a session at the Association of Corporate Counsel's annual meeting in Boston, Baffa and Lauren Leipold, another Seyfarth partner, spoke about how to move forward as AI resets expectations in the legal profession.
They were joined by two in-house lawyers, Maria Maras, associate director of IP litigation and disputes at Accenture, and Timothy Fraser, chief legal officer at Toshiba America Group.
Historically, a client has assigned work, external counsel has completed the task and the client has reviewed the bill. In the AI-enabled business model, the client and firm design workflows, and the tool and data shape delivery, according to the presentation.
"This is not to say that you're going to turn every single new intake or engagement letter into a big, drawn out negotiation," Leipold said. "What we want to do for you is just normalize the idea that you should at least be raising these questions and think through them … before you engage and get too far down the road and run into surprises later."
The session took place amid a backdrop of a growing number of lawyers using AI. Law360 Pulse's 2026 AI Survey found that the percentage of firm attorneys who said they use AI three or more times a week increased from 27% in 2025 to 47% this year. Among in-house lawyers, 56% reported that same frequency of use for this year.
The panelists on Monday urged in-house counsel to ask their firm lawyers how they use AI — or could use it — on their matters.
With AI, the work is no longer segregated between in-house and outside counsel output, Fraser said. For instance, firm attorneys will work inside their clients' platforms and vice versa. Other times, both sides will work on a shared third party's platform. As a result, a host of "thorny" questions will arise, he added.
Whose tools will be used? What are the data sources we will use to drive AI? What work should AI touch, and which matters should it not touch? What data can be used, retained, logged or excluded? How will human review and escalation work? How will each side share savings, capacity and risk?
AI also presents opportunities for a company's confidential information to be leaked, and outputs aren't always 100% reliable. As a result, both parties will want to know the protocols to deal with those issues, the panelists warned.
The lawyers on the panel acknowledged that these discussions aren't necessarily simple. For example, there likely isn't a 1:1 ratio to the amount of savings of using AI, given how much money some firms have invested in building AI tools.
"If you don't have those conversations up front, the trust that you have built up between the two of you as a working relationship is just going to completely erode because assumptions are going to be made on both sides," Leipold said. "You can't make those assumptions; you have to be having conversations so that everybody is on the same baseline."
From her in-house perspective at a technology company, Maras said lawyers should be open to experimentation, which she acknowledged can be difficult for lawyers who typically are risk-averse.
Baffa also pointed out the importance of being open to a structured trial and error process in which the firm and client work together to find a different way to price, add value and work through how they will use AI.
"It really is about creating the right kind of workflow," he said. "And that really takes a lot of collaboration up front."
--Additional reporting by Kevin Penton. Editing by Nicole Bleier.
For a reprint of this article, please contact reprints@law360.com.