Yonatan Pesses, CMO at Tailor Brands, G-CMO Forum member

What’s happening:

Once voice AI got good enough, we saw a clear opportunity. Our outsourced call center (120 to 130 reps) was slowing us down, creating quality gaps, and forcing constant retraining in a fast-moving product. We tested the tech with a small real-world pilot, and customers spent 10 to 20 minutes on calls without realizing they weren’t talking to a human. That was our signal to move.

We started narrow, with an AI IVR that routed calls, then shifted into sales. The first version flopped because we treated the AI like a new hire and fed it scripts. The turning point came when we taught it the journey instead. We gave it goals, context, and guardrails, with room to improvise like our best reps. From there, performance improved quickly, and the agents began delivering real empathy and personalization at scale.

Why this matters:

AI adoption isn’t mainly a tech project. It is an operating model shift. When you are clear about the customer experience you want and you iterate in production, AI can turn conversations into something measurable, improvable, and infinitely scalable. It reduces operational drag, improves consistency, and unlocks a new pace of experimentation.

What to try:

Start with one high-pain, well-understood workflow. Run a small pilot using off-the-shelf tools, go live quickly, and iterate weekly. Begin with a narrow role, like routing or one scenario, teach the journey not the script, and only then invest in deeper infrastructure.

The big idea: don’t wait for perfect AI. Start learning where the value is already obvious.