Before you ship the AI chatbot, just ask: Who's this really helping?

AI in CX is having its “everything is a platform” moment.

Except it isn’t an implementation problem. It’s an incentive problem.

Every vendor’s promising seamless service, happy customers, and support teams running on autopilot.

But you and I both know how this actually plays out: another over-confident bot funnel that can’t handle “refund” or “talk to a human.”

There’s a lot of deflection wearing a shiny new LLM hat. Most tools are designed to reduce support volume, not improve outcomes. And that’s a trap.

So let’s reset.

The best AI CX rollouts don’t start with bots. They start with clarity: on vision, limits, and what problem we’re solving in the first place.

Because unless you enjoy debugging LLM hallucinations in prod, you’ll want to think a little more upstream.

Here’s how I break it down when I’m talking to companies:
→ 3 strategy checks.
→ 3 reality checks.
→ And a rollout plan for teams that doesn’t set fire to your support metrics, thanks to a killer framework from Luke Costley-White (and yes it’s worth your time).

And if any of this feels uncomfortable, shoot me a message. I know how hard these calls get when the board’s watching and the burn’s real.

I’m happy to be your sounding board.