AI recruiting is software that searches for people who could do a job, contacts them, asks them the first screening questions and gathers references, before a recruiter looks at them. It is also called AI sourcing or automated sourcing. Sourcing means finding candidates who have not applied. An applicant tracking system (ATS) waits for applications. AI recruiting goes out and finds people.
The search used to be done by hand. A recruiter typed keywords such as "Android" and "Kotlin" into a profile site, read hundreds of profiles and wrote messages one by one. AI recruiting software does each of these steps with a large language model (LLM), the kind of AI model that reads and writes text. It is an example of agentic automation, in which an AI agent does the office work that a person once did.
Say a health-tech company in Hyderabad needs six Android engineers within six weeks. Only a few suitable people have applied. The recruiter types the requirement in plain words: "Android engineers with 3–6 years of experience, who have built an app used by over a lakh people, willing to work in Hyderabad."
The software searches public profiles, public code repositories and the company's own ATS, which holds people who applied for other jobs in the past. It compares each profile with the requirement by meaning, not by keywords. So it also finds an engineer whose title is "mobile developer". This method is called semantic job matching. It ranks 400 people and writes a separate message to each of the top 150, mentioning the app that person built. Forty reply. A chatbot asks each of them three screening questions: their notice period, their expected salary and whether they can move to Hyderabad. Some tools also collect a reference, an informal opinion from a former colleague.
The recruiter receives a shortlist of 18, with the reason for each match. The shortlist goes into the ATS, and the usual interviews begin. The recruiter's work is now mostly the conversations with these 18 people.
Two limits matter. India's data protection law, the DPDP Act, does not apply to personal data that a person has made public themselves, such as a public profile. But it does cover the CV, salary and other details that a candidate sends in reply. The page on data protection law explains the Act. And a model trained on past hiring can repeat past biases, so recruiters must check the shortlist themselves.
AI recruiting also has a high-volume form, blue-collar recruitment, which hires riders and warehouse workers by the thousand. Inside a company, talent intelligence uses similar models to find existing staff who could fill a role.