What AI Does in Candidate Sourcing, and Where Humans Decide

AI candidate sourcing uses software to find potential candidates from a role brief instead of asking someone to search every profile manually. It interprets the role, searches professional information, and organises possible matches into a shortlist.

The useful boundary is simple: AI expands and structures discovery; people define the hiring problem, check evidence, decide who to contact, and make the employment decision. This guide maps those hand-offs without treating a ranking as a verdict.

Talent Summoner is our product; its details were checked against the candidate sourcing, candidate ranking, and pricing pages on 19 August 2026.

What AI candidate sourcing actually does

Most AI sourcing workflows have four jobs:

Translate the brief. Turn a job description into capabilities, seniority, location, constraints, and evidence to look for.
Search beyond applicants. Find people in professional profiles and other public sources whose visible experience may fit. This is outbound discovery, not inbound CV screening.
Organise evidence. Group relevant signals, missing information, and a plain-English reason for each result.
Use corrections. Apply feedback about false positives, seniority, skills, or adjacent backgrounds to a later search.

The output is a research aid, not proof of capability or an instruction to reject candidates below a threshold.

Where humans decide

The important decisions happen at search boundaries, not after the AI produces a score.

Before the search: define relevance

Separate must-haves, preferences, and trainable gaps before seeing names. Decide which adjacent backgrounds are acceptable and remove unrelated criteria. Do not ask a tool to infer culture fit or search for a particular age or background.

After the first results: inspect evidence

Review high, middle, and low-ranked profiles. Check evidence, recency, and what the system could not verify. Missing evidence is uncertainty, not automatic disqualification. Correct the brief or weighting if results are wrong.

Before outreach: choose the conversation

A human should decide whether a candidate is appropriate to contact and what message is respectful. Verify role, location, seniority, and current employer where possible. Do not copy sensitive or speculative details into outreach. AI can suggest who deserves attention; it should not send an unreviewed mass campaign or assume consent.

During evaluation: gather new evidence

Profiles describe history, not motivation, communication, availability, or the context behind a career move. Interviews and work samples create new evidence. Keep those signals separate from the sourcing score and use a consistent process. The hiring team remains accountable for selection, rejection, accommodations, and communication; explain the decision with job-related evidence, not "the model ranked this person first".

A practical review loop for one role

Use this sequence for one role:

Write a brief with must-haves, preferences, deal-breakers, and evidence for each requirement.
Run a first search and note convincing, missing, and surprising results.
Review it with the hiring manager; mark false positives and useful adjacent profiles.
Apply feedback to a second search and compare relevance.
Verify a small group, personalise outreach, and keep interview evidence in your normal process.

This makes judgement visible and tests whether a product accepts corrections.

Sourcing versus ranking candidates you already have

Sourcing discovers people outside your applicant pile; ranking compares CVs already collected. Talent Summoner's free Candidate Ranking tool accepts up to 50 CVs in PDF, DOCX, MD, or TXT format, needs no account, and produces a shareable report. Use candidate sourcing for discovery.

Responsible use is part of the workflow

This is general information, not legal advice. In the United States, the U.S. Equal Employment Opportunity Commission's guidance on employment tests and selection procedures says selection procedures should be job-related and appropriate for their purpose. The EEOC and Department of Justice also warn that AI tools can disadvantage people with disabilities and recommend an accommodation process. Requirements differ by jurisdiction, so get local advice.

Keep a record of criteria, evidence, corrections, and who decided. Check patterns in who is surfaced or missed, and provide a human route for questions or accommodations. Public information is not automatically complete or suitable for every decision.

What Talent Summoner does and does not do

At the time of writing, Talent Summoner starts with a role brief and searches LinkedIn, GitHub, and other public professional sources across a stated 200M+ profiles. It returns a ranked shortlist with must-haves, nice-to-haves, and plain-English reasoning; feedback can adjust a later search. Sourcing starts free with no credit card.

Talent Summoner is our product, not an ATS. It does not automatically reject candidates, manage an application pipeline, or provide a contact-credit database or automated outreach sending. Your team verifies candidates, chooses outreach, runs interviews, and decides. Current options are US$39 for one Role, US$99 for three, or US$79/month for four.

Bottom line

AI candidate sourcing is most useful when it accelerates discovery and makes evidence easier to review. Set criteria first, challenge the shortlist, verify claims, and keep outreach and selection human-led. Start a sourcing search and review the results with the person who owns the hire.

FAQ

Does AI candidate sourcing replace a recruiter?

No. It can reduce profile discovery, but people define the role, verify evidence, choose outreach, assess candidates, and decide. A public profile cannot show the full hiring context.

Is an AI-generated candidate score objective?

No. It reflects the brief, available evidence, and the system's interpretation. Review reasoning, check missing evidence, and keep a human decision-maker accountable.

What is the difference between sourcing and candidate ranking?

Sourcing finds candidates outside your applicant pool. Ranking compares CVs you already have. Use the Candidate Ranking tool for that existing set.

How much does Talent Summoner cost?

The options checked on 19 August 2026 are US$39 for one Role, US$99 for three, or US$79 per month for four. Pack Roles remain valid for 24 months and monthly Roles roll for three months. See current pricing.

Talent Summoner product facts read from our live candidate sourcing, candidate ranking and pricing pages, verified 19 August 2026. We re-verify this page quarterly — tell us if something changed.