Solving the Right Problem: How Corporate Legal Teams Are Actually Scaling AI

AI is no longer a future conversation for corporate legal departments — it’s already happening. The harder question isn’t whether to adopt AI, but how to move past pilots and point solutions to build something that lasts.

That was the premise of a recent webinar hosted by HIKE2 and IntelliStack, Scaling AI in Corporate Legal: Lessons from the Teams Doing It Right. Rather than a pitch or a features tour, the session was a candid, conversation-style exchange between two people who’ve spent years thinking about how organizations actually change the way they work — not just what tools they buy. What emerged was a practical framework for finding the right problems to solve, telling the difference between optimizing today’s process and transforming tomorrow’s experience, and avoiding the traps that turn a promising pilot into a stalled initiative.

About the Session

The conversation was moderated by Jordan Dimmitt, Senior Partner Manager at IntelliStack, who leads strategic partnerships and go-to-market work across IntelliStack’s SaaS and technology alliances. Joining him were Morgan Llewellyn, Principal in HIKE2’s Innovation Practice, who brings more than 20 years of AI strategy experience along with a 2021 SaaS Product of the Year award, a 2022 AI and Automation Innovation award, and a track record implementing advanced AI for government agencies and Fortune 100 companies, and Max Brennan, VP of Corporate Counsel at IntelliStack, who leads legal across contracts, employment, IP, and compliance while running a solo in-house legal function at a PE-backed SaaS company and using AI daily to help build IntelliStack’s own CLM product.

That pairing — a strategist who has guided AI transformation across industries, and a practitioner living the reality of a lean legal team using AI to keep up — gave the session a grounded, practical honesty that’s easy to lose in more theoretical AI conversations.

Finding the Right Problem Is a Search, Not a Decision

The panel opened with a question every legal leader is quietly wrestling with: how do you make sure AI solves a real business problem instead of being adopted because everyone else is talking about it?
Morgan reframed the question itself:

“The reality is for all organizations, it’s not as simple as use AI, don’t use AI. There’s a plethora of different options within that AI sphere… This isn’t a yes or no question. This is really about search. How do you search and find the real business problem?” — Morgan Llewellyn, Principal, Innovation Practice, HIKE2

He drew a distinction between a “Chicago School” mindset focused on maximizing ROI and the “Austrian School” view of business as a search for what customers want and what an organization does uniquely well — and argued that finding the right AI problem belongs squarely in that second category. He outlined four ways organizations typically go about that search: leaning on leadership’s experience and intuition, crowdsourcing ideas internally, running scattered experiments, or leaning on vendor relationships to see where AI is already working. Of the four, he was direct about which two actually move the needle: “lean on your experience, either have a strong vision, or lean on your partners and let them basically take the brunt of the effort.”

Max’s own approach, built out of necessity as a solo legal function, reflects that same instinct. He starts by asking what work is most expensive, most error-prone, most time-consuming, and most directly tied to revenue — and for most legal teams, he’s found that question keeps landing in the same place:

“For most legal teams, the answer to that is going to involve some version of: what do our contracts actually say, and how do we get that information to the people who need it? That’s a lot of the problems that I have to solve day in, day out.” — Max Brennan, VP of Corporate Counsel, IntelliStack

From there, he said, the path is straightforward: “you get a good sense of what the problem is, then you go out to the market and find a vendor that sees the problem the same way you do” — because with AI changing as fast as it is, a shared perspective on the problem matters more than a narrow point solution today.

Optimization and Transformation Are Both Wins — But They’re Different Wins

One of the session’s clearest frameworks was the distinction Morgan drew between optimizing an existing experience and transforming it into something new. He illustrated it with an analogy:

“Concorde said, ‘Look, I’m going to keep the same experience… and we’re just going to do it faster.’ They were going to optimize on speed. And then you’ve got British Airways, who came in and said, ‘We don’t care about speed… We are going to transform the experience,’ and they created the lie-flat bed.” — Morgan Llewellyn

The question every organization has to answer honestly, he said, is “am I happy with the experience?” If yes, the work is optimization — the same experience, faster or cheaper. If not, the work is transformation, where “cost is kind of secondary” and the real question is what the experience should become.

Max’s own path traced that arc in real time. His first use of AI in contract review was pure optimization: “I was still doing the same redline-by-redline review I’d always done, but I was using AI to generate the first draft of redlines and comments. That saves me maybe two or three hours per heavy negotiation.” But those hours mattered. As a solo counsel, the time savings freed up bandwidth for the work that piles up after a contract is signed:

“Finance asks what renewal terms are. Sales asks what products they’re on. Every single time I do still have to pull up the agreement, read through it, trace an amendment chain maybe, and then email the answer back. So I’m sort of the human search engine there.” — Max Brennan

That optimization win is what’s now pushing him toward transformation — rethinking how sales and finance access contract information in the first place, rather than continuing to route every question through one person’s memory of the document. As Max put it, “the optimization wins can help inform what the transformation might look like, and also gives you a bit more time to work on it.”

Building Something That Lasts Starts With a Vision, Not a Tech Stack

Jordan asked about what comes after the initial 60 to 90 days of pilot excitement — the “stability stage” where a lot of AI initiatives quietly stall. Morgan’s answer ran against the instinct most organizations have:

“People go really deep, really fast on: what does it mean to govern these things? What does it mean to have good data? … In my experience, all those things are secondary. They almost don’t matter if you don’t have the one thing that’s absolutely necessary, which is vision. What does done look like?” — Morgan Llewellyn

He pointed to an HR technology example where the vision wasn’t a better resume parser, but to stop parsing resumes altogether and use full resumes directly in the job search process — with every subsequent decision built toward that end state. Too many organizations run a pilot without that clarity, he said, referencing Inder Sidhu’s Doing Both as a useful frame for simultaneously delivering on current obligations while still innovating.

Max’s own vision for IntelliStack reflects that same discipline:

“My ideal vision would be a world where the executed contract basically automatically populates the rest of the enterprise — pricing terms automatically generate the right invoice on the finance system side, user entitlements update in the product backend, and it starts off renewal workflows before the expiration.” — Max Brennan

Getting there requires very high trust in the semantic data pulled from contracts, since that data will directly drive business outcomes. Morgan called out Max’s approach as a model worth copying: start with the vision, then work backward to the specific data requirements that vision actually needs, rather than trying to perfect all data indiscriminately. Max added a timely reason that discipline matters even more now — rising token costs across LLM providers mean “you don’t want to spend a whole bunch of tokens boiling the ocean and have no business results at the end of it.”

The Blooper Reel: Where AI Initiatives Actually Stumble

Every framework session needs a reality check, and Jordan’s ask for “blooper reel” examples delivered one. Morgan pointed to three recurring failure patterns.

Trying to build it yourself.

“I’m going to go do all the integration… I’m going to create my own CRM,” Morgan said, describing a pattern he still sees. “You could go do that, but you’re a SaaS company or you’re a manufacturing company — is that really core to your identity?”

Becoming a one-hit wonder.

A single successful pilot with no plan for what comes next — “I rolled out my pilot, I had some success, but I don’t know what comes next” — leaves organizations without a plan for the chain of experience that has to follow.

Organizational gaps in innovation muscle.

Years of squeezing out inefficiency have, in Morgan’s words, also squeezed out the innovators: “the people who really enjoy innovation self-selected out and went elsewhere,” leaving real gaps in governance, data readiness, customer experience design, and training.

His remedy for all three: lean on vendors and partners, who can “bear some of that build burden,” suggest what comes next, and bring continuous innovation experience organizations have often lost internally. Max’s own history bears this out directly. Before AI, he built a homegrown CLM system in Jira — “it worked fine, it was not great” — and when AI first emerged, he tried a similar do-it-yourself approach through vibe coding. He hit the limits quickly:

“I can leverage them given my judgment and experience, but I can’t turn this loose on the organization and allow it to act independently of me. That just isn’t going to work.” — Max Brennan

Joining IntelliStack gave him access to engineering resources, iteration cycles, and budget no solo legal function could build alone — the clearest possible case for why mature vendor partnerships matter.

Where the Puck Is Going: Agents, Not Just Tools

Asked where they’re placing their bets on the future of AI in legal and operations, both speakers converged on the same underlying shift: from people directly operating tools and databases, to agents acting on people’s behalf.

“We used to engage with tools that then engage with databases and documents… Tomorrow, our experience is going to be very different, where all these tools are going to be interacting with agents. We’re going to be shifting the experience from this very personalized [interaction]… to my agent is now interacting with a software product on my behalf.” — Morgan Llewellyn

Morgan’s specific bet: in-house counsel will increasingly have agents reviewing RFPs and shortlisting vendors on their behalf, with law firms and AI tools themselves evaluated as “services on a shelf” — where how easily an agent can find and understand relevant information becomes a genuinely competitive factor. His advice to law firms: make sure the right information about your services is easy for an agent to discover, “because if I’m not on the shelf, if I can’t be purchased by an agent, I might be outside of the conversation.”

Max’s version of that future centers on CLM and business process automation, where AI has made it economically feasible to populate and validate data at a scale that used to require “an army of first-year associates or consultants.” His ideal end state:

“Ideal future state from my side is where I and the other folks can stay in the tools that they want to stay in, like Salesforce, have a conversation with an AI assistant that works across the org, and then the workflows happen behind the scenes.” — Max Brennan

It’s a genuinely different user experience, he acknowledged — one that “does require a lot of trust and consistency” before it’s ready to run on its own.

The Throughline

Strip away the specific examples, and the session’s message is consistent: successful AI adoption in legal isn’t about chasing the newest tool. It’s about clearly identifying the problem worth solving, being honest about whether you’re optimizing an existing experience or building a new one, defining what “done” looks like before you start, and leaning on partners who’ve already absorbed the trial and error you’d otherwise have to do yourself.

As Max put it near the close of the session, most attorneys don’t share his enthusiasm for legal ops and tooling — “I don’t want to assume that about all attorneys. And so that’s where the partnership and the vendor side is so critical.”

Let’s Talk About Your Legal AI Roadmap

HIKE2 works with corporate legal and legal ops teams to move past isolated pilots and build AI initiatives that hold up over time — from named-account governance and data foundations to the vendor and platform decisions that determine whether a promising start turns into lasting value.

If your legal team is ready to move beyond the pilot stage, we’d welcome the conversation.
Contact HIKE2 to start the conversation →