How to Measure ROI of AI in Recruitment: Metrics and a Worked Example
Recruitment is where AI ROI is easiest to measure, because hiring already produces clear data: time, cost, volume and quality. The challenge is attributing change fairly and including quality and fairness, not only speed.
Short answer
To measure the ROI of AI in recruitment, set baselines for time to hire, cost per hire, recruiter hours per hire, agency spend, candidate drop-out, offer acceptance and quality of hire, then compare after deployment. Value comes from recruiter time saved, reduced vacancy costs from faster hiring, lower agency spend and better retention of hires. Include fairness metrics such as selection rates by group, because a tool that saves money but creates adverse impact carries legal and reputational cost.
Key takeaways
- Vacancy cost, the value lost while roles are unfilled, is often the largest benefit.
- Quality of hire and early attrition protect against speed at the expense of fit.
- Fairness metrics are part of ROI, not separate from it.
- Use comparison roles or regions where possible to isolate AI's effect.
Recruitment metrics for AI ROI
| Metric | What it shows |
|---|---|
| Time to hire | Speed from application to acceptance |
| Time in each stage | Where AI removes delay |
| Recruiter hours per hire | Capacity gained |
| Cost per hire | Total recruitment cost efficiency |
| Agency spend | External cost avoided |
| Candidate drop-out rate | Experience and speed effects |
| Offer acceptance rate | Competitiveness |
| Quality of hire | Performance and 90-day or first-year retention of hires |
| Selection rates by group | Fairness and legal risk |
Estimating vacancy cost
For revenue-generating or operationally critical roles, each day a role is empty has a cost: lost output, overtime or missed revenue. A simple estimate is daily value of the role multiplied by days saved in time to hire. Agree the method with finance before launch so results are credible.
Worked example: interview scheduling agent
All figures in worked examples are illustrative, not benchmarks. Replace them with your own baseline data.
| Item | Value |
|---|---|
| Hires per year | 300 |
| Recruiter hours saved per hire on scheduling | 3 |
| Recruiter hours saved | 900 |
| Loaded recruiter hourly cost | 45 |
| Capacity value | 40,500 |
| Days removed from time to hire | 4 |
| Daily vacancy cost (average) | 50 |
| Vacancy cost avoided (300 x 4 x 50) | 60,000 |
| Total benefit | 100,500 |
| Total cost | 35,000 |
| ROI | 187% |
Attribution
- Pilot in some teams or regions and compare with similar ones.
- Track other changes such as market conditions or new job boards.
- Use before-and-after data over comparable seasons.
Why fairness belongs in ROI
If an AI tool reduces time to hire but produces adverse impact, the organisation faces legal risk, audit costs and reputational damage. Report selection rates by group alongside financial ROI. See how to audit AI hiring tools.
See ROI of AI in HR and AI in recruitment.
Related guides
- ROI of AI in HR: How to Measure, Prove and Improve the Return
Value categories, an ROI formula, full costs, a worked example and common mistakes.
- AI in Recruitment: How It Works, Benefits, Risks and Best Practice
How AI works across the hiring funnel, its benefits and risks, the law, and best practice.
- AI Recruiting Tools: Categories, Selection Criteria and How to Compare Them
The eight categories of AI recruiting tools, selection criteria and a comparison scorecard.
- How to Audit AI Hiring Tools for Bias: A Step-by-Step Guide
An eight-step bias audit method, with impact ratio calculations and regulatory requirements.
Frequently asked questions
How do you measure ROI of AI in recruitment?
Set baselines for time to hire, cost per hire, recruiter hours, agency spend, drop-out, offer acceptance and quality of hire, compare after deployment, value recruiter time and vacancy days saved, and include fairness metrics.
What is vacancy cost?
The value lost while a role is unfilled, such as lost output, overtime or missed revenue, usually estimated as daily value of the role multiplied by days vacant.
Should fairness be part of recruitment AI ROI?
Yes. Adverse impact creates legal, audit and reputational costs, so selection rates by group should be reported alongside financial returns.
How do you know improvements are due to AI?
Use pilots with comparison groups, compare equivalent periods and account for other changes such as market conditions or new sourcing channels.