AI Product Showcase - Humanly

A live product walkthrough shows how an AI sourcing platform pulls open jobs directly from an applicant tracking system, then builds a calibration set of example candidates based on who has already advanced past a hiring manager screen. That calibration list acts as a moving target: as new candidates progress in the interview process, older examples get bumped unless a recruiter pins them as a permanent benchmark.

The platform organizes candidates into three pools rather than one undifferentiated list:

• Passive candidates who have never been contacted or entered into the ATS
• A referral network built by pulling in colleagues' LinkedIn connections, turning personal networks into a searchable candidate pool without waiting for referrals to trickle in
• Existing ATS records that are refreshed and enriched over time, so a resume from years ago reflects a candidate's current role and fit

Clicking into an individual profile surfaces not just a title but the context behind it: what a candidate actually accomplished in a role, what the company did, how large it was, how fast it was growing, and what funding stage it had reached when the candidate joined. That contextual data is also searchable, not just visible.

Outreach is generated automatically by combining the job description, candidate profile data, and patterns drawn from hundreds of thousands of recruiting email threads to identify language that performs best. One example: including the word "remote" in a job description and mirroring it in outreach measurably raises response rates. A generated message can be sent as-is or edited before going out, and once a candidate replies, the system detects interest and moves toward scheduling automatically.

A live search built from a single vague prompt, "senior data scientist in Chicago," shows the system generating example profiles for a recruiter to approve or reject, refining its understanding with each response. Recruiters can also skip the conversational interface entirely and apply their own Boolean strings, company filters, and funding-stage criteria directly.

AI should never make decisions on candidates, especially applicants. It sorts and surfaces; people choose.

Throughout, matching and ranking are handled by AI, while filtering, judgment calls, and outreach approval stay with the recruiter.