The problem
Ostro Labs had product-market signal but no outbound process at all — three engineers-turned-sellers with a single spreadsheet of 850 developer and engineering-leadership contacts pulled from a community export, and no shared sense of who on that list was worth a message.
Every week spent designing a qualification process by hand was a week not spent talking to the market, which mattered more at Ostro's stage than getting the model perfectly tuned before starting.
What they configured
Ostro imported the 850-row list, kept the default 70% qualification threshold rather than spending time re-weighting it, and left company-size fit and technology relevance at the platform defaults since the team didn't yet have enough closed-won data to know what to change. The one deliberate adjustment was a modest increase to the AI / digital transformation fit weight, matching the company's own positioning.
Round-robin distribution split the queue across the three owners from day one, each capped at a daily limit low enough that a team still learning the product could keep up with every conversation it started.
“We didn't have a process to automate. We had a spreadsheet, and nine days later we had a queue and a call on the calendar.”
What happened
The queue was live the same day as the import, and the first AI-drafted openers went out within 48 hours of that — no separate qualification project, no waiting on a re-weighted model before starting. The first booked call came nine days after the initial import.
Because the scoring breakdown was stored and explainable from the first lead, Ostro's team could see which of the default nine factors were actually predicting real interest for them — data they plan to use to tune the model themselves once they have enough closed deals to trust the pattern.