Been reading a book on AI maturity lately and one idea really stuck with me, so sharing it here in case it helps someone.
A lot of us hear "we already use AI" and it's actually a great opening if you know what to listen for.
The core idea is simple. Using AI and changing how the work gets done are two completely different things.
The book splits it into three levels:
Adoption. People use AI tools in their existing tasks. Someone pastes meeting notes into ChatGPT, gets a summary, copies it back into the system. The person got faster. The process stayed the same.
Integration. AI is built into the workflow itself. Something triggers it, it does its part, a human reviews and approves. Nobody has to open a chat window and write a prompt.
Transformation. The work itself gets redesigned. Roles change, responsibilities change, sometimes the whole offer or business model changes.
There's also a second question running alongside: what value is AI actually creating for businesses? Efficiency (same work, faster)
Quality (better or more consistent work)
Innovation (doing something that wasn't possible before).
The part I found most useful is what they call the adoption illusion. Everyone uses ChatGPT, there are Copilot licenses everywhere, someone built a custom GPT, and leadership feels like they're way ahead. But if you follow the actual workflow, it's still person, manual step, AI, copy paste, system, spreadsheet, next person. AI is everywhere. The process is old.
So when someone says they already use AI, don't argue. Get curious. Something like:
"Is that mostly people using tools like ChatGPT on their own, or have you started building it into your workflows and systems too?"
And then follow the answer:
"Everyone has Copilot" = "Have you rebuilt any workflows around it?"
"We have our own GPTs" = "Does someone still have to start them and feed them manually?"
"We've automated a lot" = "Which parts still need the most hands-on work?"
That last one is the best one. You're not trying to prove their AI setup is bad. You're looking for the gap that's left after it.
Two things to keep in mind though. You can't place a company from one answer. "We have AI in our CRM" could mean almost anything, so keep asking. And higher isn't always better. Some processes are fine with adoption. Some need integration. Some are so messy they should be redesigned before anyone adds AI. And some shouldn't be automated at all, because the human judgment is the whole point.
The line I keep in my head going into these conversations:
The question is whether AI is just helping people do the old work, or whether the work itself has started to change.
It turns "we already use AI" into the start of the conversation instead of the end.
Curious if anyone else uses something similar in their discovery calls?