Before Ryan Bliss showed anyone at the Law Offices of James Scott Farrin their new legal AI platform, he'd already built them something to want.

To roll out Eve Legal, his team had interviewed staff, found their biggest bottlenecks, and created one-click tools inside Eve that solved them. 

"People couldn't get it fast enough," Bliss says. "Almost immediately they were emailing me asking, 'When can I have this?'"

Bliss runs the Workers' Compensation department at James Scott Farrin, one of the largest plaintiff firms in the Southeast, and serves as the firm's Director of Legal Technology. Getting nearly 300 employees and 60+ attorneys across five practice areas onto a new legal AI platform meant winning over people with wildly different comfort levels with technology — and different takes on whether AI belonged in their practice at all.

Here's the playbook he and Jennifer Horn, Operations Coordinator and Lead Paralegal, built to make AI stick at the firm.

Setting Up the AI Task Force

First, they assembled a five-person implementation team to spearhead the project. 

One member owned the vendor relationship to avoid the “too many cooks in the kitchen” issue, acting as what Bliss calls “the bat phone” to Eve Legal.

Each team member served as a dedicated support coordinator for one of the firm's departments, a go-to for questions, creative consultations, and problems.

They also set up a shared help desk email that routed to the entire team. "We assumed correctly that when we rolled this out, we would have lots and lots of questions," Bliss says. The shared inbox meant no one person got buried, and if someone was out or slammed, someone else picked it up. Most early questions were ones the team could answer themselves, without bouncing people to external support.

Overcommunication was built in from the start. AI platforms update constantly, and Bliss didn't want anyone waking up to a changed interface with no context. "Don't let them wake up one morning and see something new and be confused about it," he says. "Don't worry about being annoying. It's much more helpful to have your users understand what's going on."

What an Early Adopter Phase Revealed

Before any department demo, the firm ran a pilot with a small group drawn from across departments given access with one instruction: explore it and report back.

The goal wasn't just to find product friction before it hit 300 people. It was also a temperature check. The early adopters flagged that people would be worried about doing something wrong, breaking a tool they didn't understand. 

"We realized we're about to unleash this AI engine on lots and lots of employees, many of whom have no familiarity with AI at all," Bliss says.

That changed everything about how they approached training. People needed guidance on the basics: how AI works, why it sometimes gets things wrong, and why checking its output isn't optional. 

"We've all read the potential pitfalls of lawyers who utilize AI and don't double-check their output and then submit a brief and get dinged by a court because it hallucinated," Bliss says. "We couldn't have that."

Out of that came an Acceptable Use Policy, a set of standards every employee had to review and sign before getting access.

But a policy only solves half the problem. "How can we keep them from being freaked out and get them excited about it?" Bliss says. The answer was making people want access before they had it.

The Demo Strategy that Created Demand

The implementation team interviewed each department head. What are the tasks that cause the biggest bottlenecks? What do you do constantly that takes the most time? Then they built one-click tools that solved those exact problems. 

By the time the demos ran, employees weren't being asked to imagine how they'd use AI. They were shown a button that solved a problem they'd dealt with that week.

In Workers' Compensation, that meant the 30-day client check-in. Paralegals spend significant time before every call reviewing what's happened in a case, preparing talking points, and filling out templates. Bliss's team built a single button — the 30-Day Worksheet Companion — that summarizes everything that's happened over the past 30 days, generates personalized talking points, and pre-fills the template.

"When my paralegals saw it for the first time, they absolutely had that wow factor," he says. "It's such a major part of their job, and suddenly not only would it be more efficient, but it's also just a great tool to personalize that call for the client. They know exactly what talking points they have, whereas beforehand they just kind of had to make it up on the spot."

Other departments got different tools: settlement analysis for attorneys, discovery responses, case summaries. 

The Eminent Domain department had assumed AI tools were built around medical records, which their practice doesn't touch. The implementation team went to them anyway, found their actual bottlenecks, and built for those. They became the firm's heaviest Eve users.

Letting Departments Own Their Rollouts

Once the demos landed, each department designed its own implementation. The team provided structure: designate trainers, set a timeline, establish minimum usage requirements, plan for resistance — but left execution to the departments themselves.

"We went back and forth between rolling out to the whole firm versus staggering it," Horn says. "Given our size, we decided to stagger. Also, attorneys are very independent people. We decided it would be best to let each department figure out how they were going to roll it out."

The key was structure without abandonment. "We didn't want them to feel like, 'Hey, y'all have to figure this out,'" Horn says. What came back varied — the Major Injury department initially planned aggressive minimum usage requirements, then scaled back. Others went the opposite direction.

The trainer model was central to every department's plan. Each identified internal staff who would become Eve experts before their colleagues had access. They weren’t necessarily the most tech-savvy people, but trusted voices. Eve came on-site for in-person sessions with those trainers first, building competence and confidence before the broader rollout. 

By the time paralegals and attorneys got access, there was already someone in their department who'd been using it for weeks and could answer questions in plain language.

Setting Usage Floors, Not Ceilings

Every department set a minimum floor, a defined set of tasks everyone was expected to use Eve for consistently. "We want everyone to use Eve as much as possible, but here's where you start," Horn says.

The floor prevented the tool from becoming optional while leaving room for power users to go further. And the implementation team made clear that monitoring usage wasn't surveillance. It was support. "It's not Big Brother," Horn says. "It's helping everyone get used to this change."

Horn is candid about what the process actually felt like: "It's going to be this great moment, but it's going to be this roller coaster ride of things just constantly changing, changing, changing. We want to make sure everyone is comfortable with that. So we just kept saying, 'Hey, we're here to help you.'"

The Feedback Loop

Usage monitoring wasn't just about accountability. It became how the firm found and spread its own best ideas.

"Early on, I would get emails or have conversations with people who had developed these really creative, interesting use cases," Bliss says. "I got their permission to just publicize that to the whole firm."

One paralegal — someone Bliss had no idea was so into Eve — built a chatbot assistant that knows her role, her department, and her typical tasks. She uses it for everything from prioritizing discovery requests by due date to summarizing emails after returning from PTO.

"The possibilities outside of case management and work product are really exciting," he says. “It’s cool to see how people’s brains function in this new technological environment.”

The implementation team's job shifted over time from driving adoption to surfacing and spreading innovations. Best practices didn't come from the top. They emerged from people doing the work.

What They'd Tell a Firm Starting Now

Bliss starts with ethics and security. "A lot of attorneys are understandably naive, or maybe turning a blind eye to the ethical requirements that come along with using an AI tool like this," he says. "Educate yourself about how these tools work. Remain skeptical. Vet any of these companies you're going to work with."

Then: get excited and show it. "Enthusiasm for a product like this starts at the top and kind of trickles down. If you are excited and jazzed about it, your employees and coworkers will be too."

Horn's advice: AI doesn't need to be an overnight transformation. "Start with the things that annoy you the most, when you're about to go do it, you don't want to do it,” Horn adds. “If you wish you could have your own personal assistant standing beside you doing things, what would those tasks be? That's where you start."

If a firm isn't ready yet? "Getting ready is what they need to do now," Horn says. Look at your staff, your attorneys, your infrastructure. Get the interested people learning. By the time they've learned, you're probably ready to roll it out.

Read how the the Law Offices of James Scott Farrin put Eve Legal to work.