Two layers.

At AUTONOMOUS: OBSOLETE, Board of Innovation's Amir Ouki asked a room full of leaders who has the most to lose from this transition, and whether those same people are the ones deciding what changes.

Amir was recently named a partner at Board of Innovation. I work with him closely, and there's a reason. He talks about AI transformation with a clarity and conviction that's rare in this space. Direct, precise, and honest about what's actually changing.

He opened AUTONOMOUS: OBSOLETE with a question that nobody in the room was ready for.

Then he laid out a framework I've been thinking about since. Two layers of work exist inside every business. Intelligence and judgment.

Intelligence is the work of processing and producing. Research. Analysis. Drafting. Modeling. Pattern recognition. Work that can be learned and replicated. The quality of the output matters. The person who produced it does not.

Judgment is the work someone has to own. A physician who looks a patient in the eye. A portfolio manager making a call in front of a client. A CEO deciding to turn their Chief People Officer into their Chief AI Officer. Judgment requires accountability, relationship, sometimes physical presence. It can't be separated from the human delivering it.

Intelligence is the work that can be learned. Judgment is the work that has to be owned.

Wade Foster’s hackathon at Zapier gives the distinction a practical example. Zapier reports that AI adoption rose from 10% to 50% during a company-wide building week. The exercises gave people a way to try repeatable work with the tools. They still had to decide which outputs were useful and which workflows deserved to change.

But Wade choosing to put his CPO in charge of AI two years later? That's judgment. That call was entirely Wade's. It required reading the organization, understanding what the next phase actually needed, and betting on a people leader over a technologist. That decision was Wade. It couldn't be Wade's AI.

Amir's framework explains why both of those moves worked, and why they're fundamentally different kinds of value.

Take an honest look at most organizations. A big share of what people thought was judgment layer work is actually intelligence layer work. It has a judgment layer title. It charges judgment layer rates. But pull it apart and it's processes and workflows that are learnable and repeatable. Clients are starting to see that.

Three disruption mechanisms hitting at the same time

Amir mapped three disruption mechanisms. First, direct substitution: a client may be able to do a task internally that previously required an outside provider. The exposure depends on the actual workflow, the quality required and who remains accountable for the result.

Second, client self-sufficiency. Your next competitor may be your own client. If a team can produce an initial analysis or draft internally, the outside engagement may start later. That changes what the client needs to buy and what the provider needs to be good at.

Third, the erosion of the trust premium. Consulting firms, agencies, advisory shops have always charged partly because the client couldn't see inside the box. AI is making that box transparent. Clients can pressure-test a deliverable before the meeting starts and develop their own sense of what the work actually involves. When the process stops being mysterious, the pricing has to change.

All three are happening at once. Direct substitution gets the headlines. Client self-sufficiency does the most damage. Trust premium erosion is the one leaders keep avoiding.

Amir referenced an economist at Chicago Booth, Alex Imas, who's been building an argument about what happens when production gets really cheap across the board. His thesis is simple and I think it's right: scarcity doesn't disappear. It relocates.

When intelligence becomes abundant, what becomes scarce is the human element that can't be separated from the product. Presence. Provenance. Accountability. Trust. The judgment layer.

He pointed out something I liked. Look at how the wealthiest people in the world spend when they have no constraints. Live performances. Small venues. Personal relationships. Genuine human presence over the infinite content available to them. That pattern tells you where value concentrates when everything else is abundant.

So what happens to pricing? Every retainer, every hourly rate, every contract has always bundled intelligence and judgment into a single number. AI is pulling them apart. The intelligence portion reprices downward fast. But the judgment portion, once it's visible on its own, often turns out to be worth more than it was getting credit for.

That's the opportunity inside all of this. The parts of your business where a client would genuinely notice if the human were removed are becoming more valuable. Not less.

Amir also talked about entirely new service models that become viable when intelligence is cheap at scale. Continuous personalized healthcare support. Ongoing financial guidance for people who could never afford an advisor. Services that adapt to an individual in real time instead of serving the average. These categories barely exist today. They will.

Why most organizations aren't moving

And then he was honest about why most organizations aren't moving on any of this. The people with the most power to enable the transition are usually the ones with the most at stake in the current model. The partner whose book of business is intelligence layer work. The product team whose roadmap is organized around a workflow that might not exist in six months. The leadership structure built around functions that look very different a year from now.

Incentives shape behavior. And most org charts were designed in a world where analysis was expensive and human time was the bottleneck. That assumption is gone. The structure built around it is still there.

Three closing questions

He closed with three questions. I think they're the right ones to end on.

1) What share of your revenue is intelligence layer work, and what's the timeline before it reprices?

2) What parts of what you deliver are genuinely proprietary or relational?

3) What would it take to redesign around the judgment layer while protecting the revenue you're running on right now?

That last one is the hardest. And the most important.

AUTONOMOUS: OBSOLETE · Board of Innovation Opening Session · May 2026