# Screen applicants against evidence-based hiring criteria

> Read applications consistently, build a source-backed shortlist, and prepare introductions without making the hiring decision.

- Companion: Talent Scout
- Category: People & hiring
- Expected outcome: A consistent shortlist that shows exactly which evidence met each job criterion.
- Setup time: 25–40 minutes
- Cadence: On demand or twice weekly
- Execution modes: Taskilo Cloud, Self-hosted — coming soon
- Last reviewed: 2026-10-20
- Canonical: https://taskilo.ai/use-cases/ai-candidate-screening/

## What you need

- Ashby
- Greenhouse
- Google Drive
- Email

## Permissions

- Read applications for one role
- Create private review notes
- Draft candidate messages

## Safety boundary

Protected characteristics are excluded, and a human remains responsible for every hiring decision.

## Setup

1. **Define criteria.** Turn the role requirements into observable, job-related evidence.
2. **Restrict access.** Give Taskilo access only to applicants for the selected role.
3. **Audit the shortlist.** Review citations, missing evidence, and inconsistent scoring.
4. **Approve communication.** Keep all candidate messages in draft until a person approves them.

## Starter instruction

Review candidates for the selected role using only the job-related criteria I provide. Quote the application evidence that supports each criterion, identify missing evidence, and create a shortlist for human review. Ignore protected characteristics. Draft introductions only after I approve the shortlist, and do not reject or contact anyone automatically.

## Start with Taskilo

[Download Taskilo](https://taskilo.ai/download/?source=use-case&playbook=ai-candidate-screening) and configure this playbook with the permissions and approval boundary above.