Generative AI is absolutely changing the course of our lives and our work. And we all know it. For corporate legal teams, this groundbreaking technology presents a ripe opportunity to do things better, faster, cheaper—and still maintain defensibility.
It sounds like a dream come true, but getting started? Well, that takes some upfront work.
Turns out, many corporate teams are totally game.
“Historically, we’ve always seen some level of curve of adoption with new technology. More often than not, it’s us [the partner] presenting things to our corporate clients,” says Cassie Blum, senior director of review and consulting solutions at Lighthouse, a Relativity partner serving corporate clients. “Generative AI is flipping the script. We are seeing more companies come directly to us with interest and a keen eye on where they can implement it in their programs.”
Cassie was one of many experts on recent webinars all about, you guessed it, generative AI. In these discussions, the panelists answered some big questions including how do you get buy-in from stakeholders, overcome security concerns, and build workflows that work?
Here are five takeaways in-house teams can start acting on today.
Start Small to Build Confidence with AI
One key theme across the webinars was starting with manageable use cases. Large-scale adoption can be overwhelming, so it helps to begin with smaller projects where you can measure results and iterate. For Matt Bohn, senior counsel at Amgen, investigations are a great pond to dip your toe.
“Internal investigations, where the goal is to review as little as possible but have enough confidence that nothing is missed, is a ripe use case for generative AI,” he explains. “I encourage folks to use it in testing.”
Sarah Green, product marketing manager at Relativity, agreed, acknowledging that Relativity aiR for Review customers typically start with internal investigations, along with QC. “And we really encourage that,” she says. “Those use cases are where you can see immediate impact to build confidence.”
Starting small allows teams to test prompts, see AI in action on real data, and gain trust without overcommitting resources. But, as Dan Semelhack, project manager at ProSearch, another service provider proving the value of utilizing generative AI in corporate client legal workflows, points out—it’s not just the size of the project that matters; it’s the type of data.
“It’s one thing to test on the tried-and-true Enron data set, but once you see it on your own company’s documents, it’s really powerful, especially for those people who are on the fence. Once internal stakeholders see those results, they have that lightbulb moment. We’re not replacing human review—we’re validating and expediting it.”
Build Trust with Security, Transparency, and Human Validation
One of the biggest questions we receive here at Relativity is around AI’s security—and for good reason. Organizations need to understand where their data is going, how it’s being used, if it’s being stored, and myriad other considerations before bringing in a new AI solution.
“Security has been the number one thing that has been a hurdle so far,” says Dan. “At the end of the day, it’s all about comfort and making sure your infosec team feels like we’re putting these prompts, these documents, everything associated with this process, in a secure environment.”
Partnering with an AI vendor that offers transparency into their security practices can go a long way, and at Relativity we take that seriously. As Sarah shared, “Privacy and security are integral to the design of our products, and we are always willing to have more in-depth conversations. We have a ton of security professionals that will go into the nitty-gritty.”
Hallucinations also pose hesitation for those on the fence, and that’s where choosing the right AI solution comes into play. Use tools that cite sources and explain reasoning, so every AI prediction can be traced back to an original document. Relativity aiR for Review, for example, evaluates every single document and refers back to citations in the document itself when delivering its determinations so you can validate the results.
“Having the citations and rationale all right there and having the nice summary written up about why it thinks a document is responsive [is big],” says Dan. “If there’s a doc in the grey area, you’re not questioning the why. It’s all spelled out for you.”
Invest Up Front in Building AI Knowledge and Stakeholder Buy-In
Generative AI certainly has the power to make things easier, but it doesn’t come with an easy button. It works best when teams put in the work up front, building knowledge, setting expectations, and bringing the right people along for the ride. That early lift will pay off later.
“We need to shift how we think about timelines and where effort goes in,” says Cassie. “There is a knowledge building and level of effort that goes up front [with] dividends on the other side.”
EJ Bastien, senior director of discovery programs at Microsoft, agrees, saying: “People who are used to articulating a review protocol and handing it off to the contract team are going to be injected earlier into the process and be more proactive in surfacing nuances that often come up from reviewer questions.”
That means training your team (and your outside counsel) on how to use the tools and use them effectively and comfortably.
“Prompting is not a well-developed muscle, and a review memo is not equivalent to what we need for a prompt,” says Kayleigh Klinzman, program manager in Microsoft’s Litigation Group. “Figuring out how to thread the needle of what's in and what's out with natural language is harder than [outside counsel] thinks it's going to be and takes a little bit longer.”
To help, Kayleigh’s team developed quick tip sheets, reference guides, playbooks for outside counsel to help tweak prompt language in an effective way. It’s not just about documentation though; it’s also about mindset. As EJ explained, framing AI as an opportunity can help get people on board.
“Make sure everyone knows this is an exciting opportunity. Those who have been sidelined without tech chops can get into the game too, because [generative AI] uses natural language,” says EJ.
“It’s not easy” to change one’s mindset, he adds. “It takes effort, but it’s absolutely worth it.”
Broaden Your Circle
A consistent theme throughout the discussion: you don’t have to do this alone. In fact, many of the most successful teams span company walls, bringing in help from not just outside counsel but outside service providers with an expertise in the tech.
“Having vendor partners like ProSearch and Relativity helps us answer difficult questions and also gives us confidence that we’re using [generative AI] to maximize effectiveness,” says Matt. “It’s a tremendous asset to have partners to help guide us. We’re all learning together but they are certainly out in front of it and know it a lot better than most.”
Partners can help legal teams design workflows and provide a roadmap tailored to each organization’s needs. As Cassie from Lighthouse pointed out: “It’s not ‘do I know how to push the buttons’—it’s about fitting [the product] into the client’s program, and that can look different for every client. We want to walk them through each phase [… and] modulate technology to their workflow.”
Keep Learning and Stay Flexible
If there’s one guarantee with generative AI, it’s that it won’t sit still. Tools are evolving fast, and best practices are changing right alongside them. What you learned six weeks ago may already be outdated. That means legal teams need to stay curious and adaptable.
“The expertise you really need to hone is adaptability,” says EJ. “Pay attention to blogs of companies of products you use. Talk with people who are doing it more than you are doing it. Strike up relationship with someone like Cassie who sees how it’s being adopted and implemented. There are so many workflows that may not be relevant to you yet, but they will be as your structure evolves. Stay curious.”
Cassie echoes the point: “Get ChatGPT. Play with it—even outside of legal. Build that muscle memory so you can use it in other places like Copilot or aiR.”
The more curious and hands-on you are, the more ready you’ll be when the next wave of change rolls in.
Next Steps: Build Your Business Case
You might be excited and ready to adopt generative AI (and you should be!) but excitement alone isn’t enough. You need to get leadership on board, too. And we’re here to help with that.
Join us on Wednesday, September 10 at 12:00 p.m. CT for our next webinar, From Curiosity to Commitment: Building a Business Case for Generative AI in Legal.
We’ll have another great group of panelists to share their experiences and advice for justifying AI investments and aligning stakeholders across legal, IT, and procurement.
Graphics for this article were created by Natalie Andrews.
Recommended for you
