NASHVILLE — When it comes to mapping out AI use cases, many legal teams start with detailing the common day in the life of a lawyer. However, there’s one simple problem with that, notes Jamie LaMorgese, Enterprise Applications Director at law firm Holland & Hart, which has 550 attorneys across eight practice groups. "I don’t think any one day looks the same for any of them," he says.

So how do you institute a new AI tool in a complex, diverse, and sometimes confusing legal world? According to the When AI Tools Meet Legal Resistance panel at the recent ILTACON 2026 conference, law firms need to help lawyers where they already work, measure what actually matters, and focus on a singular purpose.

Meet them where they are

The acronym KISS — for Keep it simple, stupid — has long been a staple of technology design, but as it relates to legal in the AI era, it’s just as important on the adoption side of the equation. Tara Daisy, who recently left the law firm world to become Managing Director of Legal at software provider Box, said that her company's approach is to make AI as safe and approachable as possible. Practically, this means that Box lawyers can access multiple tools natively and switch models as needed, rather than forcing them to use only one tool.

“Everyone is like, my hair is on fire right now, I can’t possibly make another decision. So we try to meet you where you are,” Daisy explains. “I’m not going to force you to do something that you’re not ready to do. We’re all trying to do this on top of our day job, and we’re tired.”

How do you institute a new AI tool in a complex, diverse, and sometimes confusing legal world? According to a panel at the ILTACON 2026 conference, law firms need to help lawyers where they already work, measure what actually matters, and focus on a singular purpose.

However, the idea of “meeting people where they are” can be easier said than done, LaMorgese cautions. He reminded the audience that those testing and training on AI tools are often doing so at the expense of other work, including billable work. As a result, it’s important for legal operations and technology staff to be flexible during AI trials and evaluation, and to demonstrate the value of the time provided quickly.

“When you ask someone to do something for free, you can’t do a trial in less than eight weeks. Things pop up,” he notes. “Meeting them where they’re at is a nice thing to say, but you need something concrete.” And the idea of showing immediate value also extends to trainings, he adds. Holland & Hart has some attorneys they regularly turn to who are well respected lawyers, but also self-proclaimed technology nerds, LaMorgese says. These lawyers regularly get two weeks early access to new tools, allowing them to play around with the tool and shake down any integration issues.

“People talk about champions, but what do you do with them? Give them early access,” he explains. “They like the early access, and they talk about the tool they have with other attorneys in the firm. It builds a little bit of excitement.”

Christa Melde, AI Strategist at Thomson Reuters, agrees, noting that taking discrete steps can help legal teams achieve long-term goals better than moving into an AI future all at once. “You want to start with where you know you’re going to win," Melde says. “If you try to do everything at once, you’re going to end up disappointed... and that product is going to end upon a shelf.”

Metrics that matter

AI adoption is also helped by providing proof that AI tools are actually helping. However, the ROI question is not one that legal can answer right now. When the panel asked the audience of more than 50 attendees to raise their hand if they felt their legal team had a strong measurement for AI ROI, not a single person did.

There’s a reason for that, Daisy says: It’s near impossible to keep up with the pace of change. Legal teams don’t know how to measure success, or if they’re getting full usage of tools... or even if they need one tool or multiple.

“I think they’re not sure that they’re making the right choices," she adds. "They’re overwhelmed by the whiplash of the change. That’s the biggest thing we’re hearing — that it feels impossible to plan strategically.”

Those [lawyers] testing and training on AI tools are often doing so at the expense of other work, including billable work.

In the current AI realm, everything feels "squishy," Daisy notes. "And it’s hard to operate in that place because law isn’t squishy, it’s black and white.”

Legal teams don’t need to take this journey alone, however, Melde says. “Put your vendors to the fire about how they are able to help you calculate ROI,” she offers, adding that even if every use case doesn't provide clear ROI due to the nature of experimentation, the practice is useful. “If you can take even some of your young associates and have them put together how much it would take to do this manually, and budget in some grace for education, it should be able to help you start to calculate ROI,” she explains.

If ROI still remains elusive, the panel also pointed to ways to determine success of an AI trial before it even begins. LaMorgese says he tries to integrate his firm’s document management system and real matters into as many trials as possible. “I don’t want people to have to imagine what it’s like when it’s live," he observes. "You want as real an experience as you can get in these pilots.”

For example, one metric the firm uses is whether the technology achieves more than 70% weekly active users, which itself would be a higher proportion than the general population of lawyers. “If not, you might be purchasing shelfware.”

If adopting this sort of measurement, however, Melde reiterated that legal teams should be providing value — which means allowing lawyers in the trial to do real, tangible work with these tools. Otherwise, even the most excited lawyer in trial will fall away because they are not receiving value. “When you have those success metrics, it only happens if you can upload your firm documents," Melde says. "So whatever that looks like, make sure it’s baked into your evaluation process.”

The definition of success

As the panel discussed, success isn't always determined by the solution itself, but by how it's applied within the legal team.

Daisy says she starts any trial by asking, "What are your end goals?” And to get there, it may mean re-evaluating processes rather than technology. “We ask, ‘This is how we do things today. Should we? No, that doesn’t make sense anymore.’ And people sometimes get surprised by it," she explains.

AI adoption is also helped by providing proof that AI tools are actually helping. However, the ROI question is not one that legal is easily able to answer right now.

Other times, however, success comes down to getting everyone aligned on the goal. LaMorgese relayed the story of the 1995 New Zealand sailing team in the America's Cup. When preparing for the race, team leader Peter Blake responded to every strategy or build question with: “Does it make the boat faster?” That became the guiding light that helped New Zealand take the trophy away from the US team for only the second time in more than a century.

“When you talk about all the features you need or don’t need, [leaders] are really concentrated on the IT portion of AI,” LaMorgese says. “But we don’t know what race we’re winning a lot of the time.”

Indeed, Melde adds, it’s important for legal teams to remember the ultimate purpose of AI implementation is to help lawyers do their jobs. “You’re not buying a feature,” she says. “You’re buying an efficiency tool. You’re buying something to make lawyers better.”

You can find more coverage of ILTACON from this year and past years here