VIDEO: AI Talent Sourcing: Technology Roles

Reading about autonomous sourcing is one thing. Watching it run — from a blank project to qualified, ranked candidates — is another. This replay walks through exactly that on a real technology role.

What happens in the video

The run starts with a new project. Upload the job description and the hiring manager’s intake notes — including the details that never make it into a JD, like the role being on-site in Austin, Texas — and Calyflow reads both to build the plan.

From there the agents take over:

• Sourcing Plan — where to look, drafted from the JD and notes.
• Qualification criteria — a matrix of testable signals (Java, microservices, architecture) used to score and rank every candidate.
• Sourcing sessions — the agent runs against a small starting budget, finds qualified matches, drafts outreach, and stores results in Calyflow. It checks what’s already been sourced, so it never surfaces the same person twice.

By the end of the demo, two newly qualified candidates land on the shortlist, then more are saved as the run continues — all while the cost holds steady at roughly 31 to 41 cents. Two candidates for about a dollar, with the shortlist showing the quantified qualification metrics behind each rank.

Why it matters

This is the difference between chatting with AI and running a workflow. The criteria come from the attached JD, not from whatever someone typed that afternoon, so every candidate is scored the same way and every result carries the evidence that put it on the list. Run it again next week and you get the same caliber of output — for the price of a coffee, rounding error.

Want to run your own search? Create a free account — your own API key, no credit card.