Screen applications against written role criteria
Applications are read against the must-have criteria the hiring manager wrote down, and each candidate gets a short, sourced summary of what the resume shows. Recruiters see a sorted queue with reasons and decide who moves on. For teams that receive more applications than they can read.
Function: HR · Industry: Any industry
Screen applications against written role criteria
6 steps
At a glance
- Trigger
- A candidate applies to an open role.
- Typical inputs
- Resumes and application answers
- Role description with must-have and nice-to-have criteria
- Hiring manager notes
- Job-specific knockout questions
- Systems it usually connects to
- Greenhouse
- Lever
- Gmail
- Slack
- How the work splits
4 rules, 1 model, 1 person steps
How it runs, step by step
6 steps, each tagged by what does the work. Rules are fixed logic, a model reads or drafts, and a person decides.
- Rules Fixed logic in code
- Model AI reads or drafts, with a confidence score
- Person A named person decides
- 1
Apply knockout rules
RulesLegally required answers, such as work authorization and location, are checked by fixed rules written by the recruiter.
- 2
Summarize against the criteria
ModelA model lists, for each written criterion, the evidence in the resume and quotes the line it came from, or says it found none.
- 3
Sort the queue
RulesCandidates are ordered by how many written criteria have evidence, and every candidate stays visible.
- 4
Decide who moves forward
PersonA recruiter reads the summaries and the resumes, and chooses who gets a screen. Nobody is rejected by the system alone.
- 5
Send the status message
RulesCandidates receive the confirmation or decline message the recruiter chose, on your timeline.
- 6
Audit the process
RulesThe criteria used, the summaries and the decisions are logged so you can review outcomes by stage.
Where people approve
The decisions that stay with your team, whatever the confidence score says.
- A recruiter makes every advance or reject decision.
- The hiring manager signs off on the criteria before screening starts.
- Your HR lead reviews outcome data across stages on a regular schedule.
KPIs to track
What to measure, before and after. Targets come from your own baseline.
- 01Time from application to recruiter decision
- 02Recruiter hours per hire at the screening stage
- 03Candidates advanced who passed the first interview
- 04Pass-through rates by stage, reviewed for unintended patterns
Questions about this workflow
Is automated resume screening a fairness risk?
It can be, which is why this design avoids ranking people by a model score. The model summarizes evidence against criteria you wrote, a person decides, and you log the outcomes by stage. Several jurisdictions regulate automated hiring tools, so we check the rules that apply to you with your counsel before launch.
Can it infer things the resume does not say?
No. Criteria are matched to stated evidence, and a missing item shows as missing. That keeps the summary checkable and avoids guesses about seniority, age or background.
Which steps use rules, a model and a person?
Rules: apply knockout rules; sort the queue; send the status message; audit the process. Model: summarize against the criteria. Person: decide who moves forward. Rules are fixed logic in code. A model reads or drafts and reports its confidence. A named person decides.
What do we need to start?
Typically: resumes and application answers; role description with must-have and nice-to-have criteria; hiring manager notes; job-specific knockout questions. You also need read access to Greenhouse, Lever, Gmail and Slack, or the equivalents you use. A workflow audit measures your volumes and past cases first, then gives a fixed price for the build, so you can decide with real numbers.