> ## Documentation Index
> Fetch the complete documentation index at: https://docs.jobin.cloud/llms.txt
> Use this file to discover all available pages before exploring further.

# How to shortlist and review candidates in AI Search

> Keep AI Search results under control: shortlist the people you want, set aside the rest, and always see what is left to review.

To shortlist candidates in AI Search, open the **To review** view, shortlist the profiles you like, and mark the others **Not a fit**. Live counts show how many are left.

## What this gets you

* You always know what's decided and what isn't: candidates to review, shortlisted, and not a fit each have their own view with live counts.
* You never review the same person twice. Rejected profiles move out of the way, and reopening a search returns you to what's undecided.
* Your shortlist is ready to act on: select it and start outreach.

## Steps

1. Open your search in **AI search**. It opens on the candidates still awaiting your decision.
2. Work through the **To review** view one candidate at a time.
3. Choose to shortlist a candidate you want to pursue, or mark them **Not a fit**.
4. Check the live counts for candidates to review, shortlisted, and all results.
5. Open **Shortlisted** to review the people you've kept. Rejected profiles stay in their own view.
6. When **Refresh Order** appears, choose it to reorder results using the latest scores.

## Tips

* Jobin shows a completion state once you've reviewed every available candidate.
* Switch between reviewed candidates and the complete result set at any time.

## Ideas to get more from it

* Review in short sessions between calls. Jobin remembers where you stopped.
* Use the completion state as your finish line, so you know when a search is fully reviewed.
* Run [AI screening](/how-to/screen-candidates-with-ai.md) on the shortlist for a closer comparison.

## Related

* [AI Search guide](/ai-search.md)
* [How to find candidates with AI Search](/how-to/find-candidates-with-ai-search.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.