Explore the pilotof recruiters use AI to speed up hiring
The problem
Hiring is not short of applications.
It is short of signal.
One-click applications and generative tools have made applying faster and more polished. Employers still have to find the credible evidence hidden inside the volume.
A real high-volume example
1,057applications in five daysThis is not a sourcing problem.
It is an evidence problem.
The market evidence
Employers want efficiency.
Candidates want accountability.
CV-Library's April 2026 research describes a widening trust gap: AI is now common in recruitment, but employers and candidates remain concerned about strong people being screened out without meaningful human judgement.
say a lack of human intuition can mean missing top talent
say strong applications are filtered out before interview
of candidates believe AI rejected them without human review
have abandoned or considered abandoning an AI-led application
Source: CV-Library, AI in hiring is filtering out strong candidates, April 2026.View source
The workload, quantified
A first pass can consume an entire working week.
Even before interviews, correspondence, scheduling or checks, a single role can create tens of hours of basic application review.
At 1 minute each17.6hours
At 3 minutes each52.9hours
And that is only the first screening pass.
Calculated from 1,057 applications. Figures illustrate review time, not an industry-wide average.Illustrative direct review cost
£528–£2,381At an assumed loaded reviewer cost of £30–£45 per hour, before any interviews or hiring-manager review.What the numbers conceal
The real loss is not just time.
Volume overwhelms judgement
When hundreds of applications arrive in days, careful review becomes an operational burden. Speed begins to replace attention.
Presentation is mistaken for proof
Early screening often rewards polished wording and keyword alignment before anyone has tested whether the claims can be substantiated.
Strong candidates become invisible
Relevant experience can be buried by volume, non-standard career paths and applications written for people rather than screening systems.
Trust erodes on both sides
Employers inherit unexplained scores. Candidates experience silence or rejection without knowing whether a person meaningfully considered them.
Why the current stack breaks
Most early screening measures presentation—not proof.
Existing tools solve useful parts of recruitment. The gap is a consistent evidence screen between application and interview.
ATS keywords
Fast sorting, but it can reward CV optimisation and familiar phrasing rather than demonstrated ability.
Knockout questions
Useful for basic eligibility, but answers remain self-declared and context is easily lost.
AI detection
It targets how an application may have been written—not whether the applicant can support the claims.
Checks and interviews
Stronger evidence usually arrives late, after scarce interview time has already been committed.
The missing layer
Ask for comparable evidence before spending scarce interview time.
InterventionAI creates that evidence layer while keeping a trained, accountable person in the loop.
See the solution