A good AI sourcing review makes the search logic visible. Start with the brief, compare the result with known evidence, check the source boundary and record the next human decision. This worksheet gives an agency team a shared way to do that.
This worksheet is for executive search firms and agency recruiters who want a reviewable process before they refine a search or contact a new person. It treats every result as a working hypothesis. A ranking supports review; it does not replace the recruiter’s assessment.
Start with a shared search record
Save the exact brief, date, locale, source scope and the first result set. If the search is rerun later, the team can see what changed. A clear record separates role definition, database context, source coverage and review standard.
Use four labels for every important criterion:
- Required evidence: a requirement that needs a visible, relevant signal before the profile moves forward.
- Strong signal: useful evidence that increases confidence but needs interpretation.
- Preference: a helpful characteristic that can be negotiated or tested later.
- Needs clarification: information the profile does not establish and the recruiter must verify before a consequential decision.
This distinction matters because an unverified point is a question, not an automatic exclusion. It also stops a vague preference from quietly becoming a hard requirement.
The four-pass search review
1. Test the brief for hidden exclusions
Rewrite the role as outcomes and observable evidence. “Commercially minded VP” is a direction, not a test. “Owned a pricing change across two markets and reported the result” gives the sourcing team something to look for. Add adjacent titles and career paths only when they are plausible for this assignment.
Ask a subject-matter reviewer to mark each line as hard, negotiable or unclear. Do not change the search while you are still discovering what the client meant. If the brief contains a location, language, sector and exact title requirement, run a separate check on whether all four are genuinely mandatory.
2. Inspect evidence, not just the match label
For the first profiles, write down the actual signal: a role, responsibility, product, market, credential or public work example. Then add one sentence about what still needs clarification. This creates an evidence trail that a partner or client can review.
Yena’s sourcing workflow is designed to show why a profile may fit rather than reduce the review to a keyword hit. Use that explanation as a prompt for human checking, not as an automated recommendation. The Yena sourcing page describes the product boundary: search and ranking support the researcher; the recruiter remains responsible for relevance and outreach.
3. Separate source coverage from search quality
A search should be reviewed alongside source coverage. Check the internal CRM first, then the approved external sources available to your team. Compare the result with two or three known relevant profiles. Record whether each profile is present in the source, present in the CRM or represented under a different identity or title.
Do not fill the gap by collecting every possible field. GDPR Article 5 requires personal data to be adequate, relevant and limited to what is necessary for the purpose. Your review log should capture the source, purpose, date and decision needed—not become an unrestricted second database. Read the official GDPR text on EUR-Lex before turning a search experiment into a new data-collection process.
4. Decide whether the next action is search, review or stop
Use three clear next actions:
- Search again: one agreed criterion changes, and the new test is recorded.
- Review evidence: profiles require a human check, a client question or a lawful contact decision.
- Pause: the team records what evidence or permission is needed before continuing.
A documented pause protects the candidate relationship and gives the client a clear next question instead of a padded shortlist.
Copy this search-review worksheet
| Review field | Write down | Decision |
|---|---|---|
| Assignment version | Date, team contact, market and role outcome | Which brief is the team using? |
| Required evidence | Criteria that need a visible, relevant signal | Keep, clarify or negotiate |
| Evidence signals | What would support each requirement? | Visible, partial or needs clarification |
| Known reference profiles | Two or three relevant people to test coverage | Found, absent or misclassified |
| Source boundary | CRM, public web and approved providers used | Enough coverage for this market? |
| Next action | One change, one review or a pause reason | Team contact and date for the next check |
Keep the worksheet beside the search record. A recruiting CRM is most useful when the reasoning travels with the candidate and the assignment, rather than disappearing into a private spreadsheet. Yena’s recruiting CRM workflow is built around candidate history, notes, outreach context and next steps; use those fields to preserve decisions, not to imply certainty.
What human oversight looks like in practice
Human review should be a real intervention point. The EU AI Act sets human-oversight requirements for high-risk systems in its scope; whether a particular recruiting use falls within that scope depends on the system and context. The official AI Act text on EUR-Lex is the right place to check the legal wording; this article is not legal advice.
For a sourcing team, that means a reviewer can explain why a profile was kept, referred for clarification or paused. A candidate is not contacted simply because a system produced a high-looking match. If the source, purpose or contact route needs clarification, the responsible decision is to pause.
Run the worksheet on one real assignment
Choose one open role and run the review with the recruiter, a subject-matter colleague and the client contact. Keep the test small. Compare the original result with the result after one agreed change. Note what improved, what needs clarification and which assumptions should return to the client.
Then repeat the exercise during Yena’s 10-day free trial using a real workflow. Compare the current plans, request a demo or connect the review to the client hand-off workflow when the team is ready.
Frequently asked questions
How should a recruiter review an AI sourcing search?
Review four elements in order: the role brief, the evidence expected, the source coverage and the data or privacy boundary. Change one agreed assumption at a time, record the result and decide whether the next step is another search, human review or a documented pause.
What does a small result set tell a recruiting team?
It is a prompt to inspect the market, brief, records and source coverage. Compare the result with known relevant profiles and document which requirement or evidence gap explains each inclusion, exclusion or unresolved case before changing the search.
Should recruiters broaden every search that returns too few candidates?
Broaden only a criterion that the client has agreed is negotiable. Keep requirements, evidence gaps and preferences separate; a larger list is useful only when the added profiles still have a clear reason for review.
What part of the review can Yena support?
Yena can search the public web, approved data providers and a recruiting CRM, then help organise ranked profiles and visible matching signals. Recruiters still review the evidence, context, contact route and lawful basis before deciding what happens next.