The problem
A residential real-estate agency competed against more than a thousand rival firms and many times that number of independent agents. Then the market itself contracted by roughly two-thirds, and the demographic targeting everyone leaned on stopped working.
The agency already believed in jobs-based segmentation — but applying it meant analyzing conversations, and it was fielding 150-plus calls a day, 30 to 40 minutes each. One person doing that by hand was physically impossible. The methodology that could have saved them simply didn't scale.
Letting AI find the jobs at scale
The team built an AI pipeline that transcribed every call and scored it against Advanced Jobs-To-Be-Done in about three minutes. Run across roughly 15,000 conversations, it collapsed 185 tiny micro-segments into four core jobs:
- Investment — preserving capital. Just 30% of leads, but two-thirds of revenue.
- First apartment — driven by fear of a costly mistake; the buyer wants the agent as a mentor.
- Upgrading living space — more room, tangled up with anxiety about moving.
- For the family — buying toward a "safe future."
The AI also flagged which phrases moved deal probability up or down, so agents could read intent live instead of guessing.
What changed
The jobs view reshaped the whole operation:
- Sales. Agents were trained and scripted per segment, and messaging shifted from square footage to stability and a safe future — the language buyers actually used.
- New services. A trade-in offer was introduced specifically for the upgrade segment, removing the moving anxiety that stalled those deals.
- Content. An AI-assisted content system generated segment-matched posts, cutting the copywriting and design headcount to near zero while increasing output by about 1.5×.
The results
Even as the surrounding market fell by two-thirds, the agency held its ground and improved its core funnel:
- Agent-to-sale conversion rose about 20%.
- Conversion from social posts to qualified leads rose about 40%, thanks to job-matched messaging.
- Monthly revenue held flat in a market that had shrunk dramatically — effectively a large relative gain.
The takeaway: jobs-based segmentation was always the right idea — automation is what made it usable on every single call.