Can we use gender predictions to balance a shortlist?

No. Finowl gender predictions must never be used to screen, filter, rank or select candidates, including attempts to balance a shortlist. Use them only for anonymized, aggregate outreach analytics.

How should we treat unknown results?

Report them as their own category. Unknown is a normal result when a name is ambiguous or not found, and forcing a guess would distort the aggregate picture.

Should we use AI mode for recruiting analytics?

Usually not. Batch items default to dataset-only lookups, which cost 1 credit. AI mode costs 3 credits per lookup and adds little value for aggregate reporting on real names.

Gender prediction · HR and recruiting

Aggregate outreach analytics for recruiting teams

Talent acquisition teams often want to know whether a job advertisement, sourcing channel or employer brand campaign reaches a broad mix of people. Where self-reported data is unavailable, Finowl gender predictions can estimate the aggregate composition of an anonymized audience, such as everyone who clicked through a campaign in a quarter. That helps a team see whether its outreach, not its selection, is narrowing the funnel.

Finowl gender predictions must never be used to screen, filter, rank, score or make any decision about a candidate or employee. Run analysis on anonymized data, separated from applicant tracking records, and report only aggregates. A prediction estimates how a name is typically used, not how someone identifies, and it is wrong for some people. Many jurisdictions restrict processing of this kind of data in employment, so involve legal counsel.

Try it in the sandbox About gender

USE CASES

Where it fits in hr and recruiting

Job ad channel reach
Estimate the aggregate gender mix of anonymized visitors from each advertising channel to see whether a job board or social campaign reaches a narrower audience than intended, then adjust where ads run.
Employer brand campaign review
Compare aggregate estimates before and after rewriting job ad language or imagery, to understand whether outreach changes broaden who engages, without connecting results to any applicant.
Event and webinar outreach
For recruiting events and webinars, estimate the overall mix of anonymized registrants to evaluate whether invitations, partner communities and promotion channels reach a varied audience rather than the same familiar networks.
Diversity of outreach reporting
Produce quarterly, aggregate-only reports on outreach reach for leadership, clearly labeled as statistical estimates and kept strictly separate from hiring metrics, applicant tracking data and every individual record.

HOW TO INTEGRATE

The workflow

1
Export first names from anonymized outreach data, such as campaign registrants, with no link back to applicant tracking records.
2
Assign opaque ids and send up to 50 items per call to POST /v1/gender/batch.
3
Aggregate identified results by channel or campaign, counting unknown results separately rather than guessing.
4
Delete per-record predictions immediately after aggregation and never write them to a candidate or employee profile.
5
Share only aggregate reports, with a note that figures are statistical estimates about outreach, not about hiring decisions.

POST /v1/gender/batch

curl https://api.finowl.co/v1/gender/batch \
  -H "Authorization: Bearer $FINOWL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "items": [
    {
      "id": "reg-7f3a",
      "type": "name",
      "value": "Jean",
      "country": "FR"
    },
    {
      "id": "reg-91c2",
      "type": "name",
      "value": "Priya",
      "country": "IN"
    },
    {
      "id": "reg-c04d",
      "type": "name",
      "value": "Kim",
      "country": "US"
    }
  ]
}'

Estimates gender for anonymized webinar registrants to report aggregate outreach reach by channel. Results are deleted after aggregation and are never used to screen, rank or decide anything about a candidate.

GOOD PRACTICE

Before you ship

Never use predictions to screen, rank, shortlist, score or make decisions about candidates or employees. Doing so is harmful, likely unlawful in many places, and outside acceptable use.
Keep analysis anonymized and physically separate from applicant tracking and HR systems, document it in your privacy notice, and seek legal review before running any employment-related analytics.
Where you collect demographics, ask people to self-identify voluntarily with inclusive options. Self-reported data is preferable and should always replace any estimate in reporting.

Read the API reference for every field and error code. Each lookup costs 1 credit (3 when it requests AI); see pricing.

A FEW GOOD QUESTIONS

Common questions