3 Talent Acquisition AI Agents, Live in Production.

Job postings shape the applicant pool more than any sourcing channel does, and they are usually written by copying the last one. That carries forward phrasing accumulated over years — aggressive framing, unnecessary years-of-experience floors, credentials that are conventional rather than required — each of which narrows who applies without narrowing to anyone more capable. Omitting the salary does the same thing more sharply, filtering for candidates who can afford to apply speculatively. On the CV side, parsing tools extract everything available, including details that should never inform a screening decision, and then infer the fields that are missing — an inference that is a guess presented as data.

These agents constrain what enters the process and what leaves it. Job Posting Builder turns a role brief into a posting that describes the actual job, states the salary, and drops the wording that quietly narrows who applies. Resume Parsing pulls structured fields out of a CV without inferring the ones that are not there, and strips the details that should never reach a screening decision — the refusal to infer being as important as the redaction, because an inferred field is indistinguishable from a stated one downstream. Offer Management checks an offer before it goes out against three things: the band, what colleagues in the same role are actually paid, and what was agreed at interview. The second is the check almost nobody runs, and it is the one that prevents a new starter arriving above a longer-serving colleague doing identical work. Nothing is sent.