Eight unscripted sessions produced 59,706 words of transcript in a single half day. Legal leaders from across the industry talked through AI, resourcing, pricing, and the future of the profession, and five terms rose above the rest.
Taken one at a time they look like separate concerns, the sort of thing each session might have raised on its own. Followed closely, they turn out to be one argument told from five angles: a single shift in what legal work is becoming, and how it will get resourced, priced, and trusted.
Judgment
The most-used substantive term, and the reason is structural rather than rhetorical. As AI compresses a large volume of work into a manageable set, the value of a lawyer’s time concentrates in the decisions that remain. The center of the role is moving toward those decisions, which is a sharper claim than saying AI clears away the routine and leaves everything valuable in place. High-stakes, high-risk, and complex matters will keep demanding human judgment even as the routine volume around them thins out, which means the work that survives is the work that was always hardest to systematize.
The premise underneath everything else is that legal work is coming apart into an execution layer and a judgment layer. Once that separation is real, the unit of legal value stops being the output and becomes the decision. That reframing changes what a legal team buys, what it measures, and what it is willing to pay for, and every other term here is downstream of it. If judgment is the thing being priced and protected, the questions of who supplies it, what accelerates it, and how it earns trust stop being separate conversations and become one.
“Your lawyers become the architects of judgment and not just the producers of answers… far more strategic and way more scalable than spending a career revising the same clause for the sixth time.” — Rishi Varma, Cargill
Flex talent
Once the decision is the unit of value, the question of who is positioned to make it follows directly. Flex talent belongs in the conversation as a standing resourcing option measured against the alternatives, not as a stopgap for a headcount gap. The framing that matters starts a level up, with why a legal team buys anything at all: why it carries internal headcount, why it engages outside counsel, why it uses an alternative provider, and why it leverages flex talent. Each answers a different need, and flex talent’s answer is proximity.
Outside counsel engagements are discrete and hyper-specialized, and those lawyers do not sit with the business. Flexible resources learn how the company makes money and come to understand what moves the needle culturally as well as legally. That closeness is the point, because judgment that is not grounded in the business is judgment about the wrong things. The objection that has always capped this approach is spin-up cost, the worry that new resources burn expensive time learning the business before they produce anything. That objection is weakening, and the reason it is weakening connects directly to the next term. When context transfers faster, the cost of bringing in the right person drops, and the case for keeping every matter with whoever holds it now gets harder to make. Judgment sets what the work is worth, and faster context transfer settles who gets to do it.
“A massive benefit to flex talent is just the proximity to the business that they can come in and embrace.” — Aaron Boersma, Ford
Agents and agentic
What makes context transfer faster is what these two terms describe, the same change seen from different altitudes. At the level of a single role, paralegal work is moving toward legal agentic engineering, where the job is less about producing a document and more about directing the systems that produce it. At enterprise scale, agents get embedded in the triage and routing of thousands of requests, grounded in internal playbooks and institutional expertise, working as a genuine force multiplier rather than a novelty.
There is a significant gap between what the technology can do and what people are actually doing with it, and pretending otherwise helps no one. But the direction is not in doubt. Look at what the supervising job actually involves. The person running the agents is no longer producing the work; the work has become deciding whether what the agents produce is right and fit to hand over. The skill that matters becomes knowing whether the output is right, which is another way of describing judgment. The AI shift and the judgment shift run as one trend rather than two parallel ones, because the more execution the agents absorb, the more the human role compresses toward the decision.
“We’re going to need people who have that technical capability and understand how these agents work… and what their potential risks are.” — Mike Haven, Meta
Trust
That connection is why trust is the connective tissue between the AI story and the outside counsel story. Adoption is a function of trust, and the newest form of it is agentic trust: the governance, visibility, and transparency that let a client hand real work to an agent with confidence. You cannot delegate to a system you cannot see into, and you cannot delegate to a person whose judgment you have not learned to rely on. The same requirement sits underneath both.
The requirement runs through pricing as directly as it runs through technology. If AI changes a firm’s cost structure and the firm cannot be transparent about how that flows into fees and measurable client outcomes, the relationship has a problem no tool solves. Trust does double duty in this industry, serving as the oldest thing in the client relationship and the newest requirement of the technology at the same time.
Anyone can report what happened; the trusted advice is what to do about it and how to feel about it, and that judgment layer is exactly what firms should hold onto as they lean on AI to deliver the rest. Trust is also increasingly the moat. As more work is turned over to machines, the ability to translate expertise into a human relationship matters more, not less, and the firms that earn the next engagement will be the ones talking about how they partner to deliver value rather than only about rates or which tool they have bought.
“Ultimately adoption is going to be a function of trust.” — Sabastian Niles, Salesforce
The billable hour
The one term that surfaces everywhere. Conversations about AI adoption, data infrastructure, flexible careers, and panel strategy all find their way to it, usually without anyone planning to raise it. Ask whether hourly billing stays dominant in five years and the answers split, and the split is more useful than a consensus would have been. The model holds for strategic expertise and judgment even as it falls away on more commoditized work, and it persists for structural reasons too, since panels are still built around hourly rates and standing up unit-cost service lines takes years of investment before it can carry real volume.
But the deeper pressure may not come from where everyone expects. Clients have never managed to force the billable hour out, and the more likely path is that firms retire it themselves once the math stops working. With fewer hours underneath the pyramid, a firm cannot bill its way to growth, and what is left to price is judgment. That inverts the usual story, which has always assumed alternative pricing arrives when buyers demand it. The economics of the firm look set to force it first, once the leverage model that made the billable hour work stops producing enough hours to sell. When that happens, the thing left to charge for is the same thing every other term pointed at.
“The billable hour is like the Twinkie of the legal world. It’s gonna survive nuclear winters.” — Eric Dodson Greenberg, Cox Media Group
One system, not five terms
If the execution layer is increasingly handled by agents and the value concentrates in the judgment layer, the old pricing model runs out of hours to sell, and the industry has to learn to price the decision instead of the time.
Judgment is becoming the unit of value. Agents are absorbing the execution that judgment used to be buried inside. Flex talent is the resourcing model that gets judgment closest to the business fastest. Trust is what lets work be delegated to both agents and new people. And the billable hour is the pricing model caught in the middle, holding where judgment lives and eroding everywhere else.
These five terms trace one shift rather than five separate concerns, a legal function reorganizing around the decision, and the question ahead is not whether it happens but how fast, and who builds the infrastructure to make deciding, resourcing, and trusting work together.
All eight FlexFest sessions are available on demand. Watch here.
