AI is already delivering measurable value in legal work – but not everywhere, and not equally, either. How can you isolate the wheat from the chaff when you’re looking to optimize and fine-tune your AI strategy?
According to many 2026 AI Visionaries, experimentation (done properly!) and collaboration are key. From accelerating data analysis to improving consistency and surfacing key insights faster, many of them are fully immersed in the biggest benefits legal AI has to offer. At the same time, they emphasize that AI has its limits: human oversight, iteration, and client alignment remain critical.
Let’s take a look at striking the right balance to ensure you’re getting all the ROI you can from your AI investments.
Don’t: Adopt AI Just Because Everyone Else Is Doing It
Don’t get us wrong: you should be adopting AI. The point, though, isn’t just to check a box, say “we’ve got it!” and soldier on. Intentional choice, fit for purpose, and iteration are essential to the successful and defensible use of AI.
“The inertia of existing systems is a powerful force in the public sector. This can be a good thing. Change just for the sake of change is not a good policy for institutions that rely on public trust, so any substantial tech improvements must be built on solid footing,” says Dave Fronapfel, an e-discovery professional with a nonprofit organization. “There are no shortcuts.”
Thoughtful implementations of AI must balance your team’s and clients’ sentiments, skills, use cases, and risk appetites, bringing all this detail into evaluation and selection processes to ensure no due diligence stone goes unturned.
“Our job as AI evangelists is to educate about strengths as well as pitfalls. That means that we can’t just ignore the weaknesses or risks that this new tech brings. Admitting that there are situations where generative AI and LLMs are not the best solution will help to engender trust,” Dave notes. “One of my mantras all along has been that generative AI is not a carte-blanche replacement for any of our existing workflows – it is simply another tool in our kit. Ensuring that we utilize it deliberately and responsibility is crucial in this industry where defensibility is paramount.”
For Ray Mangum, a partner at Redgrave, this big-picture view applies to every individuals’ AI habits, too.
“You have to do more than play around with AI. You've got to try to solve a real problem. You have to struggle with it and overcome those struggles. That's how you build up an internal sense of what AI can and can't do well,” Ray says.
Do: Explore Lots of Use Cases for AI
Fortunately, there is plenty of room for that experimentation. Potential use cases for AI abound, as do your options for reliable, principled legal AI systems that can help tackle them.
A hallmark that’s true across the board for AI Visionaries? They’re trying all sorts of things, learning fast, and adapting based on those lessons – all the time.
Dave, speaking from a highly technical role, says: “I am utilizing generative AI to create code to help automate tasks that may have taken me days to write in prior years, but which can now be done in minutes. The ability to quickly respond to unique data problems is really impactful.
But, he notes, using AI doesn’t have to mean jumping right into coding projects.
“AI is also invaluable for fact finding. Its ability to understand natural language queries means less reliance on keywords and opens up so many potential unique use cases,” he explains, highlighting the storytelling and strategic work many legal practitioners do every day. “The fact that there is software to perform such an abstract task as comparing testimony to find inconsistencies is an example of how amazing this technology can be for an attorney to have in their toolkit.”
“AI cuts through data to highlight the key information quicker than ever. This allows senior practitioners to focus on the higher-value work of doing something with that key information quickly,” Gary Foster, managing director at Alvarez & Marsal UK, agrees. “In e-discovery specifically, we see the value of human AI experts interacting with the AI models – and that’s where many of our project success stories have come from.”
For Gary’s team, those success stories are fantastically diverse: “We have an AI-powered tool that analyzes image files and design drawings, which we use when working with our construction clients. We are also exploring agentic AI and are already having some success with AI agents acting as junior investigators within organizations and, although still in early stages of development, we’re looking at how AI agents can supplement teams of investigators. For example, AI’s ability to understand all languages may mean companies can deprioritise language skills when they look to build their internal teams.”
Ray, too, emphasizes the value of AI in legal data intelligence work like e-discovery.
“Where my practice really focuses – large-scale discovery – generative AI is game-changing. We're seeing outstanding results in terms of recall and precision, and it's performing well even in complex reviews like internal investigations,” he adds. “Those results don't happen on their own. There's real judgment and skill involved: some prompts work much better than others, and you need to understand the legal landscape to know what's defensible, what should be disclosed, and what needs to be negotiated with opposing parties. You also need to understand workflows: where gen AI works best, how it fits with other workstreams, and where traditional approaches are still preferable. And it's always changing.”
Again, balance is key. But if you get it just right, you’ll strike productivity gold.
“AI has transformed how I work. Working with AI assistants doesn't just make things easier and faster – when done correctly, it produces better results,” Ray says. “The key is iteration: going back and forth, staying engaged, remaining in charge of the thought process. You have to understand and own the results.”
Don’t: Rest on Your Laurels
Successful AI adoption isn’t a one-time exercise or a single-run race. Once you’ve onboarded a fantastic tool or optimized an AI-powered workflow, the work keeps right on rolling.
“There are many ethical obligations legal technology professionals have when it comes to building, training, or using artificial intelligence in their work. The key is privacy, transparency, and fairness,” Gary explains. These are ongoing conversations: with software providers, colleagues, clients, judges, and all of your stakeholders.
Defensibility is a critical driver of this work, of course. But so is competition – because it’s hotter than ever.
“Clients are going to expect more, faster. And they're going to expect work product that doesn't contain errors AI should have caught,” Ray predicts. That means legal teams will need to keep up with evolving technology that’s, every day, more and more capable of meeting those demands.
“AI is going to become mainstream, and managing AI will be a core professional skill,” he adds.
“Continuous education is not optional – I feel like there is some new development on a weekly basis that impacts e-discovery,” Dave agrees. “This industry was always tough on those who rest on their laurels, but I feel this aspect is only accelerating since new innovations are seemingly coming at a firehose pace.”
Do: Learn from the Experts – and Share Your Knowledge
Thankfully, because everybody is talking about legal AI, there are plenty of resources out there to help you keep up with that pace.
“To be at the forefront of using AI, it is key to surround yourself with all different kinds of AI experts and to be in a place that allows you the flexibility to react when you find the opportunities to do so,” Gary advises.
“AI enthusiasts are going to have to stay on top of developments, which are coming fast,” Ray concurs. “I check Hacker News multiple times a day, and I read blogs aimed at technologists rather than lawyers – Simon Willison's blog, for example.”
Actively participating in the legal data intelligence community can help, too. So many brilliant people build their careers in this space; most of them are enthusiastic in conversation and eager to share what they’ve learned. Attending industry conferences, for instance, is a great way to make these connections while earning some CLEs, too.
Ultimately, you’ll want to give back what you learn: to peers, internal stakeholders, and clients. Learning about AI and how to discuss it from the experts will help you defend, advocate, and communicate about it more effectively in the settings where it matters most.
“Being able to show that you understand a client’s concerns, have taken steps to address them thoughtfully, and can explain why those measures are adequate is often the key to resolving hesitations about AI,” explains Ray. “In other cases, it's critical to be honest that use of AI can involve tradeoffs that must be balanced by the client.”
As the saying goes, a rising tide lifts all boats: the more of these savvy conversations happen around the legal world, the more our profession is empowered to thrive in the AI era.
Graphics for this article were created by Kael Rose.
