When GenAI arrived in late-2022, law firms found they needed to paint a new picture with a blank canvas, and a recent ILTACON panel explored how a few firms have worked to make their own GenAI masterpiece
NASHVILLE — As the world approaches the two-year mark since the original introduction of OpenAI’s ChatGPT, law firms already have made in-roads into establishing generative artificial intelligence (GenAI) as a part of their firms. Whether for document and correspondence drafting, summarization of meetings and contracts, legal research, or for back-office capabilities, firms have been playing around with a number of use cases to see where the technology may fit into the future.
At the same time, however, there’s a recognition that not all GenAI implementations are created equal. And given the nascent nature of the technology, coupled with increased interest from law firm leadership who may be itching to push up adoption timelines, law firm technology teams saw in GenAI a project unlike anything they’ve seen before. The world of GenAI implementation has been one of shifting priorities, flexible goals, and a continual need for feedback and communication to make sure firms stay on the right track.
It’s akin to painting on the fly, noted the Designed for Success: Mastering the Art of Artificial Intelligence panel, at the 2024 International Legal Technology Association Conference (ILTACON). Yet, with careful consideration, the right brushes, and a whole lot of patience, some firms are creating their own GenAI masterpiece. Here’s how three large law firms approached their own blank canvas.
The initial design
When developing BakerHostetler’s GenAI program, Katherine Lowry, the firm’s Chief Information Officer and head of its IncuBaker tech subsidiary, said she may have been starting from a blank canvas in terms of the technology itself, but not in terms of people and process. Due to the firm starting IncuBaker in 2016, she already had a team she felt “could have iterative moments of looking at the technology and seeing if it’s a fit,” which helped when GenAI arrived.
Given their leg up, they were able to more quickly plug the technology into a seven-step innovation framework: Discover; Explore; Manage risk; Share knowledge; Experimentation; Monitor; and Release.
Given the nascent nature of the technology, coupled with increased interest from law firm leadership who may be itching to push up adoption timeline, law firm technology teams saw in GenAI a project unlike anything they’ve seen before.
Vishal Agnihotri, Senior Director of Knowledge and Innovation at Alston & Bird, said that the discover and explore phases were a bit different for this technology in particular. Compared to past technologies, Agnihotri said, “it feels like this one had the most hype or excitement attached with it.” As a result, rather than trying to convince the firm to use the technology, she found she had the opposite problem: “We were not saying it was magic, there is a lot of hard work,” Agnihotri explained. “It’s a very interesting thing to work around when you’re used to pitching technologies for years in your firm.”
At Crowell & Moring, Chief Innovation and Value Officer Alma Asay said the firm tackled this problem by encouraging experimentation — in an ethical way. The firm used GenAI early for “legal-adjacent” work that did not contain client information, and then that evolved into experimentation within Microsoft’s Azure system, which allowed for more contained use case development.
During this period, Asay added, leadership meetings commonly revolved around the question: “What technologies can we use safely so we can start to push the envelope, and which do we need to be more cautious about?”
Build vs. buy
With the discovery and exploration period firmly underway, the ILTACON panel noted that in the middle of 2023, discussions started to move towards implementation and adoption — or more crucially, the decision whether to build their own technology or purchase it from outside providers.
Asay says that she was in the buy-mindset upon joining the firm in 2021, especially as the former founder of a legal technology company herself. However, she found that with GenAI, the tools were so new and there were few options for what the firm wanted to do, that “we didn’t want to spend speculative cash” in hopes that the tools would catch up.
Instead, Crowell leaned on Microsoft’s tools on the back end and utilized low-code development to build their own technology, rather than slow down development. “It was really just that sense we wanted to keep pushing forward,” she said.
At Alston & Bird, Agnihotri said the firm wanted to take a multi-pronged approach. First, similar to Crowell, they built with an Azure and OpenAI back-end, which “will get us warmed up and trained enough to understand how to do this.” Second, the firm “didn’t want to try and emulate content powerhouses,” so for use cases requiring large amounts of legal data, it would “evaluate those tools on their merits.” Finally, firm leadership looked at buying legal use cases from Microsoft itself — a new option, given how Microsoft’s tools can be adopted directly to legal use cases easier than ever before.
“This is different from any other technology that we have explored or rolled out, because the dust is still settling on all this.”
Of course, this decision will be different at every firm, and notably all three firms on the panel are global in scope. No matter the size of firm, however, it’s worth exploring where it might fall on Gartner’s buy-versus-build spectrum in regard to GenAI, said BakerHostetler’s Lowry, estimating that her “strategic mindset would be that we’re going to be 80% to 90% consumption.”
She also found that with GenAI, there were more opportunities to embed and extend technologies further than she would have anticipated. For example, when talking with GenAI vendors, she realized “we could build as fast as they could build.” However, even if a firm can build GenAI tools, that’s not always the answer: “We can build that. Do we want to? Maybe not.”
Picking and choosing spots
The panel agreed that GenAI will have a crucial place in the law firm of the future. Exactly where that place is, however, remains to be seen. “This is different from any other technology that we have explored or rolled out, because the dust is still settling on all this,” explained Agnihotri, adding that her firm has received mixed client signals regarding GenAI.
In some cases, she said, attorneys would receive RFPs from the firm’s business development team that mentioned AI, along with outside counsel guidelines restricting its use from the same client on the same day. Ultimately, rather than rolling out GenAI to every client matter, Agnihotri explained that “we’re always looking for a signal from the client for interest.”
Crowell’s Asay agreed, joking that she has a four-year-old that is really into the game, The floor is lava. “And I feel like I’m playing that all day, every day, with generative AI.”
Even internally, she said, she realized she was at risk of losing people’s trust with a scattered rollout. The firm started using ChatGPT, then rolled out its own proprietary Crowell AI soon after, then mandated training on Crowell AI or users would lose the access they already had. She worried attorneys might not appreciate all the external forces mandating these rapid changes.
“I had to be very open with people and say, I know we’re asking a lot of you, it’s the nature of what we’re doing here,” she noted. “And I was really amazed with the willingness of attorneys to roll with it.”
Agnihotri added that these stops and starts aren’t likely to change soon, even as GenAI as a whole maintains momentum forward. She said Alston & Bird adopted a mantra familiar to those in the northern United States and Canada: Drive carefully in snowy and icy conditions, because that’s when GenAI is today.
“You never hit the brakes,” she said. “Gen AI’s journey was you can never hit the brakes, because that’s not an option, but you can’t go full speed ahead either.”
You can find more articles from the ILTACON events here.
