One framework per role. Every side of every candidate.
A good hire only comes from looking at everything that matters for the role, every time, and in the same way. AI Talent Pro builds a framework for each role and holds every applicant to it. It works from evidence, does the same checks for every applicant, and records every result.
Six layers,
one standard.
Each layer answers one question about a candidate, and you can open each one to see its working.
Understand the role
The framework starts from what you know: the job description, the hiring manager's notes, the intake call and, if you have one, a "what good looks like" CV. AI drafts it; you review, edit and save it before it goes live.
Separate the requirements
Must-haves, weighted categories and knockouts are kept apart, because they do different jobs. Your agreed requirements take priority over loose wording in the advert. So a stray phrase can't rule anyone out.
Judge on evidence only
Each category is scored on its own, with the evidence that supports it. If the evidence isn't there, it counts as missing. Nothing is guessed or filled in.
Keep the maths ours
The AI judges the evidence; the platform does the arithmetic from your weights. Knockouts follow the rule you chose for each one: limit the score, or simply flag it.
Learn from conversations
Interview and call transcripts add new evidence, and the next analysis takes them into account. A topic the conversation didn't cover is neutral. Candidates are never marked down for something nobody asked about.
Record and re-check
Every assessment is stored with its inputs and reasoning. When you change what candidates are screened against, the platform offers to re-screen those already screened. If a CV is removed, any score built on it is withdrawn.
It reads experience the way a specialist would.
Keyword filters reward CVs that use the advert's words. The framework looks at what the person has actually done. It weighs experience by the work itself, not only by job titles or years in one industry. So someone who has done the job in a neighbouring sector isn't missed.
- ✓Two passes: a fast screen of every applicant as they arrive, then a detailed analysis of the shortlist
- ✓The same standard for the first applicant and the two-hundredth
- ✓Strengths, gaps and what to probe in interview, not just a number
Honest about what
it couldn't check.
A framework is only as good as its weakest moment. These rules make sure it never quietly gets something wrong.
No CV, no score
If there isn't enough CV text to judge, you're told. A candidate we couldn't check never looks like a poor fit.
Failures shown as failures
If an assessment fails, it says so and can be run again. It never appears as a low score.
Fair by instruction
Only role-relevant criteria are scored. Protected characteristics are explicitly ignored.
Score changes explained
When a score moves, for example after a call, you see the old score, the new one and the evidence that moved it.
People decide
The framework ranks and explains. Your team shortlists, rejects and hires.
Your standard stays yours
Each role's framework belongs to your company and is never shared with other customers.
About the framework.
Can I see exactly how a candidate was scored?+
Do you publish the exact method?+
What happens if we change the criteria part-way through?+
Does the AI make the hiring decision?+
Put your next role through the framework.
Start your 14-day free trial and see every applicant assessed against your own standard. Full system access, no card required.