Candidate matrix and mandate matrix
Candidate evidence and mandate requirements are structured independently, so neither is distorted by the other and both are reusable.
All features
Matching & Matrices
A match score you can put in front of a client.
Most AI matching is a black box with a percentage on it — impossible to defend when a client asks why candidate three ranked above candidate one. RayCruit's core score is computed by fixed rules across 12 dimensions and 90 criteria — not by a language model: reproducible, inspectable and identical every time you run it.
- 12 dimensions, 90 criteria, rule-based scoring
- Gates evaluated before any averaging
- Requirement, value, score and reason on every criterion
- Unspecified requirements marked, never silently penalised
The problem
A percentage without a reason is worse than no percentage at all.
Opaque matching creates a specific failure: the recruiter cannot tell whether an 82% means a strong candidate with one gap or a mediocre candidate with no disqualifiers. Worse, a hard requirement — the right to work in the country, a mandatory licence — gets averaged in with everything else, so a candidate who cannot legally take the job still scores respectably.
- Black-box scores cannot be explained to a client or challenged internally
- Hard requirements get averaged away instead of blocking
- Missing information is silently treated as a negative signal
- The same candidate scores differently on different days, so nobody trusts the number
Two matrices, compared by the same rules every time.
RayCruit maintains candidate intelligence and mandate intelligence separately, then compares them. Separating the two is what allows one candidate assessment to serve every open mandate — and what makes the comparison auditable.
Rule-based scoring
The core score is computed across 12 dimensions and 90 criteria. The same inputs always produce the same output — which is what makes it comparable across a shortlist.
Gates before scoring
Hard requirements such as work authorisation are evaluated as gates before any scoring happens. A failed gate is a failed gate, not a lower average.
Evidence on every criterion
Each line shows the mandate's requirement, the candidate's value, the score and a readable reason — so the ranking can be inspected line by line.
Explicit 'not specified'
When a mandate never stated a requirement, the matrix says so rather than penalising a candidate for failing to satisfy something nobody asked for.
Staleness detection
Change the CV or the mandate and the matrix flags itself for regeneration, so a shortlist is never ranked on last month's information.
Optional AI analyses
Where genuine judgement helps rather than computation, dedicated analyses cover risk, positioning and interview preparation — each grounded in the structured facts.
How a match is produced
Nothing in the sequence is hidden, and every step is inspectable afterwards.
- 01
Structure both sides
The candidate matrix and the mandate matrix are built from validated structured data.
- 02
Evaluate the gates
Hard requirements are checked first. A failure is reported as a failure, not folded into a score.
- 03
Score the criteria
Remaining criteria are scored across the 12 dimensions, each with its evidence and confidence.
- 04
Review the matrix
The recruiter reads criterion by criterion, and adds judgement where the data cannot decide.
- 05
Explain it to the client
The same structure becomes the fit explanation that goes out with the submission.
How search firms use it
Ranking a longlist
Forty candidates ranked consistently against the same structured brief, with the reasoning attached to each position.
Defending a shortlist
When a client asks why a candidate was included, the answer is a criterion-level comparison rather than a recruiter's impression.
Finding the real blocker
A strong candidate failing a single gate is visible immediately, which turns a rejection into a solvable question.
What changes
Defensible recommendations
Client conversations move from assertion to evidence.
Faster longlist triage
Consistent scoring removes the slowest, least reliable part of manual comparison.
Trustworthy numbers
Determinism means the team can rely on the score instead of second-guessing it.
See it in motion
The platform, end to end, in two minutes.
How the modules hand over to each other in the real product — customer and mandate, candidate and matrix, the workspace where a pairing is decided, and the material that goes to the client.
2 min
Continue across the platform
Placement work does not stop at one module. These parts of RayCruit connect directly to this one.
Candidate Intelligence
Structured CVs, skills, languages, eligibility and a reusable candidate matrix.
Learn morePlacement Workspace
The candidate × mandate workbench for evaluation, process, feedback and delivery.
Learn moreAI Orchestration & Compliance
Every output checked against a fixed structure, versioned prompts, provider logging and staleness control.
Learn moreSee the modules working together.
A demo follows one mandate across every module, which is the only way the connections become obvious.