A few times a year, I get to facilitate a huddle where our members, CEOs of mid-market companies across a wide range of industries, come together to compare notes on whatever's keeping them up at night. Last week the topic was AI.

Two of our members who have successfully implemented AI initiatives across their companies, Tom Stirling of Stirling Brandworks and Joel Sitak of Foley, joined to share their experience. Combined with the questions and experiences of others it led to a robust conversation about AI in companies today and common themes for the future.

We held a huddle on this same topic about a year ago. The difference was obvious within the first few minutes. A year ago, people were experimenting personally and comparing tools. This year, nobody was asking whether they should be doing something with AI anymore. They're asking whether they've chosen the right two or three things to focus on, and whether they're building them well and with the right tools.

Read the business headlines and you'll see why. AI adoption is now the top priority CEOs name for 2026, ahead of revenue growth and ahead of hiring, according to SHRM's research heading into the year. And a World Economic Forum survey of global CEOs found that roughly half believe their own job stability now depends on getting AI right. And when AI is constantly evolving, it's hard to get it right.

Read enough of that research and a picture forms fast: everyone is moving, nobody is fully sure they're moving in the right direction, and everyone is feeling the uncertainty.

Is Your Company's AI Use Single-Player or Multi-Player

Almost every CEO in the group already uses AI somewhere. The split we discussed was between people using it individually and people who've turned it into a shared capability across their team.

Tom put a name to this: single-player versus multi-player. Most organizations, his included until recently, have a bunch of people getting individual value out of AI on their own laptops, their own prompts, their own workflows, with almost none of it shared or built into how the company operates. He spent the last six months building the "multiplayer" version for his agency: a shared system with agents that draw on one central set of client and company context, so the work doesn't depend on who happens to be prompting.

Joel's company has gone down that same road. They've made decisions about and centralized their AI architecture, given it real engineering ownership, and built agents that touch go-to-market, engineering, and customer success, all pulling from the same underlying data.

Here's the pattern worth naming: the CEOs getting real value aren't the ones with the most sophisticated individual AI habits. They're the ones who stopped treating AI as a personal productivity trick and started treating it as infrastructure.

Stop Trying to Build the Perfect Database

Joel shared that no one has ever successfully built one perfect, unified database, and trying to is a trap. His company uses a knowledge graph instead, a system that can recognize that "Tom," "Thomas," and "Tommy" in three different systems are the same person, without requiring every system to be cleaned up and merged first.

That idea resonated with others. One member, running a manufacturing company with an ERP, a CRM, and a stack of spreadsheets, described feeling paralyzed trying to build a data lake before he even knew what problem he wanted AI to solve. Joel's advice was to name the two or three high-value problems first, then figure out what data those specific problems actually require. Don't wait for the golden record. It's not going to happen.

That is advice today for AI, but it's longstanding advice for how CEOs approach any hard capability gap. Start with the destination in mind, not with an inventory of everything you might eventually need.

‍The Harder Conversation Is About People

The research backs up something I heard directly from our members: this comes down to trust as much as it does technology. Nationally, close to 40 percent of employees say AI is making them worried about their own future at the company, and the CEOs leaning hardest into AI are often the ones most aware of what that displacement could mean for their own people.

Joel shared that his company is tighter in some functions now because of AI, and he talked openly about the need to treat leaders as stewards of both people and agents going forward. Tom's version of this was about earned trust: his shift toward recording nearly every client and internal conversation wasn't universally welcomed at first, but it stuck once people saw it actually saving them time rather than just watching them. Joel also helps his call team to understand that the AI evaluation of their calls helps them to grow and improve in their jobs.

The CEOs who seem to be handling this most thoughtfully are the ones being straight with their teams before the ambiguity turns into fear.

What CEOs Can Take From This

The value of that hour and a half came from two members willing to show their work, and from the questions everyone else brought to the table. That's what a peer group of CEOs actually doing this work can offer that a report or a webinar or a synthetic board can't.

A Harvard Business Review article published in March 2026 supports this. Researchers found that when AI was asked for strategic advice, there was a consistent pull toward whatever's currently trendy, almost regardless of a company's actual situation.

Judgment built from lived experience is something no model can replace, and that's exactly what Tom and Joel offered when they stood up and shared what actually happened inside their companies.

If you're still deciding where to start: pick the two or three problems where AI could actually change your business. Build shared infrastructure before you build more individual habits. And have the people conversation early, honestly, and before your team hears it from someone else.

___________________________

‍

Learn more about Katahdin Group’s CEO Collective and Key Executive Collective, rigorous process-driven peer groups focused on getting you to the next level.

‍