Five live segments, six weeks, no 80-page deck.

The challenge

The product had a few thousand active customers and growing churn that nobody could explain on a slide. Sales pitched expansion the same way to a 10-seat customer as to a 200-seat one. CS prioritized accounts by gut feel and account-manager workload, not by signal. Marketing wanted lifecycle campaigns that respected where a customer actually was, not just plan tier.

Leadership had been promised segmentation studies before. They came back as 80-page PDFs that nobody used. The ask this time: ship segments that live in the CRM and change how the GTM team works on Monday morning.

The approach

Six-week Rapid AI Pilot, scoped to behavior signals — usage, adoption depth, engagement patterns, support load — rather than demographics or firmographics alone.

• Week 1 — Discovery. Interviewed sales, CS, and product. Aligned on the decisions segmentation needed to support: who to call, what to pitch, when to escalate.
• Weeks 2–3 — Modeling. Built behavior-based segments on 18 months of product usage and engagement data. Tested cluster stability across time slices.
• Weeks 4–5 — Activation. Pushed segment assignments to the CRM as a refreshing field. Co-wrote the play-by-segment playbook with sales and CS.
• Week 6 — Handoff. Trained the GTM analyst to retrain the model and adjust segments quarterly. Mapped the 90-day path to layer in churn prediction next.

What we shipped

• Five behavior-based customer segments, named in language the GTM team uses.
• Segment assignment pushed into the CRM as a refreshing field — visible on every account record.
• A play-by-segment playbook covering sales motion, CS cadence, and lifecycle messaging.
• A lifecycle reporting framework the GTM team adopted as its standing weekly view.
• A 90-day roadmap to layer in churn prediction on top of the same data foundation.

Outcomes

Why it worked

Behavior beat demographics. The segments cut across plan tier and company size — and that's exactly why they predicted what the GTM team needed to do, not just what the customer looked like. The other unlock: co-writing the playbook with sales and CS in the room. The model was 60% of the work; the playbook was 40%, and it's what made adoption stick.

Could this work for your team?

If you have 12+ months of usage data, a GTM team that wants to act on segments (not just see them in a deck), and a willingness to put model output into the CRM, this is the shape of engagement that fits. See Targeted Analytics for Growth or the Rapid AI Pilot for next steps.