The problem
A B2B SaaS platform uses AI to scan thousands of public group chats, score each message for buying intent, and surface ready-to-contact leads — with the contact and a link back to the original message. Its edge over keyword-based rivals is context: it understands what a message means, so customers don't drown in false positives. Buyers range from staffing and recruiting agencies to consultants, accountants, and law firms.
But the product had grown by accretion — a long, unprioritized backlog, features added on a hunch, and no system for deciding what mattered next. Worse, almost nobody activated on their own. New users couldn't tell whether the product would actually deliver enough leads to be worth paying for, so every deal leaned on a manual presales conversation. Self-serve activation sat at 2%, and the founder was the bottleneck for all of it.
The job hiding behind the demo
About twenty interviews — fifteen with real users — reframed the picture:
- The core job was a steadier flow of inbound clients. Adjacent jobs kept surfacing too: finding partners and widening a professional network.
- Many users didn't even know the product could notify them — they were burning whole days monitoring chats by hand.
- The decision blocker was concrete: before paying, buyers needed to see how many relevant leads they could realistically expect each month for their niche.
What changed
The team rebuilt onboarding around proving value before the paywall, with an AI agent doing the work a salesperson used to:
- The agent asks 3-4 plain-language questions, then instantly shows real leads matched to the user's business.
- The user picks 5-15 that look right; the system reveals the contacts — a live demonstration that the channel produces.
- Only then does it offer the upgrade. The whole path collapsed from seven-plus steps to four.
Behind the scenes, the team adopted a scoring formula to prioritize the backlog and rewrote external messaging in the language of the job — inbound client flow — rather than feature lists.
The results
- Self-serve activations grew 7.5×, from 2% to 15% — new users now register and pay entirely on their own.
- Revenue rose about 30%.
- The manual presales bottleneck disappeared, freeing the founder from hand-holding every signup.
The lever wasn't a bigger sales team — it was letting the product prove its value before asking for the credit card.