AI in the Real World: Takeaways from EPI Denver’s AI Panel

AI pays off for owner-led businesses when it sits on top of a sound revenue system, and it creates expensive messes when it doesn’t. That was the thread running through the AI panel our Managing Partner, Charlie Warden, joined this month for the Denver chapter of the Exit Planning Institute.

Charlie shared the stage with Alison McQuillan of The Streamliners and Tony DiDonato, who runs a fractional IT firm in Denver. Robert Lee moderated. Robert opened by noting that everyone in the room had been pitched AI in the past month, probably more than once, and promised this hour would be different: three operators who work inside small and mid-sized businesses every day, talking plainly about what works, what’s overhyped, and what has to be in place first.

Charlie took the more pro-AI seat on the panel, given we use AI every day across client engagements and inside our own firm, and we’ve also seen how quickly it goes wrong without strategy and expertise behind it.

What we said on stage

Strategy comes before tools. When we evaluate a growth strategy for a client, AI doesn’t change the fundamentals. Brand, marketing, sales, and customer success still need to run from one strategy on solid revenue operations. Once those pillars line up, you can ask where AI adds efficiency and where it supports your people. Skip that step and you get customer support running a homemade GPT agent on someone’s personal account while marketing ships AI-written content legal never saw. AI lets a company build a mess faster than any tool before it, Salesforce included.

Revenue Operations is where the return shows up. The loudest promises, like “we can replace your sales team,” are the ones falling flattest. Our team has tested dozens of AI tools, and few add real enterprise value in sales or content on their own. The payoff we see is in RevOps: connecting teams, systems, and processes. Put the right people and systems in place, add automation where it fits, then layer agents on top to make decisions and route work.

Expertise is the ingredient. We shared two examples from our own work. We built a development agent that now handles websites, SEO, and AI visibility better than we could alone, after roughly 200 hours training it on principles our team developed over 20 combined years. We also helped incubate a telecom brand inside a larger company that passed $1M in ARR within 12 months with three full-time people supported by agents. That result rested on the brand, secure data infrastructure, and governance built over the prior three years, plus a phased plan for the agentic team.

Vet your partners hard. Charlie has talked with around 50 AI implementation shops, and most don’t hold up. Look for real strategic depth and a software development background, because the discipline of building, testing, and deploying software still applies. Watch for firms that wrap ChatGPT or Claude, mark up usage 3x, and sell it back to you as something you’ll never own. Alison added a sharp point: shops without technical depth often reach for an agent when simple automation would do the job for a fraction of the cost.

Start with the unglamorous basics. For a 10 to 50 person company, the highest-value work is a clear AI policy, a properly secured team account, real time spent training it, and guardrails and SOPs for every user. That gives you a low-risk test bed. Tony’s advice to start with repetitive tasks and Alison’s AI maturity model pointed the same way. Our addition: the ability to deploy something overnight is no reason to do it.

What it means for owners planning an exit

Buyers still ask the same questions. One attendee asked whether acquirers give credit for AI investment. From what we’ve seen, private equity still focuses on margins, efficiency, and whether the business can scale with 10 people instead of 12. AI can help you get there, and buyers will notice the results in the numbers. On the venture side, we expect today’s AI bubble to consolidate, and more VCs are asking for real revenue as early as seed.

Human channels are gaining value. Wally Waldron asked about attribution and where growth channels are heading. Our view is that in-person, community-driven channels will grow more valuable, even though they’re hard to attribute. As privacy rules limit tracking and AI content labeling takes shape, human storytelling becomes a stronger growth driver. A room of advisors talking face to face made the case well.

Thank you

Thanks to Robert, Wally, and the EPI Denver chapter for hosting, and to Alison and Tony for a candid conversation. Watch the full recording here.

One offer from the panel still stands. If someone pitches you an AI tool and you can’t tell whether it’s real, send it to us. We’re happy to take a look.