We don’t have an AI adoption problem anymore. We have an AI results problem.
Agencies are using AI. ChatGPT. Copilot. AI inside the AMS and CRM. Saved prompts. New tools. Employees experimenting.
We’re doing a lot of AI.
But ask a harder question: What has it actually changed?
Did your team save meaningful time? Did producers close more business? Did the customer experience get better?
Stop measuring AI activity. Start measuring AI outcomes.
Three ideas stood out.
I heard the phrase AI-maxxing at UNBOUND, and it describes where many agencies are right now.
Another tool. Another prompt. Another feature. Another experiment.
Experimenting is good. But more AI isn’t the goal.
Instead of asking, “How can we use more AI?” ask: “What do we need to get better at?”
Pick four or five problems that matter. Decide what success looks like. Put a number next to it.
Then figure out where AI fits.
Outcomes over optics.
This may have been my biggest takeaway.
Two agencies can use the same AI tool and get completely different results.
Why?
Context.
Your customers. Your workflows. Your data. Your processes. Your people. Your goals.
AI needs to understand your agency to become more useful.
So maybe the question isn’t “Which AI should we buy?”
Maybe it’s: “What does our AI need to know?”
Give AI generic information, and you’ll get generic answers.
Your context is the advantage.
AI can write an email in seconds.
Your employee knows when not to send that email.
They know which customer needs a phone call. Where work gets stuck. Which process makes no sense. What isn’t written in the procedure manual.
That’s valuable context.
One idea I liked from UNBOUND: Have employees create an instruction manual for themselves.
How do I work? What am I responsible for? What information do I need? What does good work look like?
Give AI that context.
Then give it something hard.
Pick one difficult problem and work on it with AI every week.
Outcome. Context. People.
That’s what I’m taking away from UNBOUND.
Outcome: What are we trying to improve?
Context: What does AI need to know?
People: How can our team use it to do better work?
You don’t need more AI for the sake of more AI.
You need better results.