Decision support · advisory only

Hire on the whole picture,
not the loudest data point.

Today, a candidate’s personality test, cognitive test, and skills test arrive as three separate PDFs from three separate systems. Managers compare them by eye — or give up and go with gut feel. UTIE pulls the three together into one clear picture, so the decision rests on the candidate, not the paperwork.

Try it nowHow it works

The problem today

Fragmented reports

Behaviour, cognition, and skills live in three different systems. Nothing joins them. Comparing across is manual, slow, and easy to skip.

Gut-feel takes over

When the data is hard to read, managers fall back on intuition. The science gets ignored and bias creeps in.

Numbers without certainty

A score looks identical whether it came from one stale report or three fresh ones. How much to trust it is invisible.

What UTIE does

One profile, six dimensions

Architectural foresight, execution resilience, stakeholder influence and three more — each tied to a real on-the-job outcome.

Confidence on every score

How much data backs each number is drawn straight on the chart, so you know what is solid and what is a guess.

Advisory, never automatic

Every output is decision support, never the decision. A human stays in the loop, with full provenance behind every score.

Try it with a real report

How it works

From three PDFs to one trustworthy profile, in four steps.

The engine is deliberately not a black box. Every number on the dashboard is the result of these four steps — and the math is open in the repo.

01

Extract

Each report (OPQ32, Verify G+, AMCAT) is parsed by a strategy that knows its layout. We pull the raw scores and capture date — and refuse files that aren't machine-readable, rather than fabricating numbers.

02

Normalise

Different tests speak different languages. OPQ uses STEN (1–10, bell-curved); Verify G+ and AMCAT use percentiles. We convert every score onto the same percentile scale so they can be compared and combined fairly.

03

Infer composites

Six composite dimensions are computed — each a weighted blend of constructs across all three sources, tied to a real on-the-job outcome. If inputs are missing, the weights renormalise and the confidence band widens, never fabricating signal.

04

Score confidence

Every output ships with a confidence number built from four signals: how much of the expected data is present, how reliable the source tests are, how recent the reports are, and how much the overlapping scores agree with each other.

The six composites it computes

Each composite was fitted against a real performance criterion — the column on the right is what the score is predicting, not a slogan.

6 dimensions
  • Architectural foresightSupervisor-rated design quality
  • Execution resilienceOn-time delivery & defect rate
  • Stakeholder influence360° peer influence rating
  • Analytical problem solvingLive coding & case-study performance
  • Adaptive learningTime-to-productivity on a new stack
  • Delivery ownershipManager-rated accountability
Try it now

Drop a real OPQ32, Verify G+, or AMCAT PDF — the radar updates live.

Try it now

Drop a report. Watch the profile build itself.

One PDF per source on the left. The radar on the right updates as the engine reads each report, normalises the scores onto a common scale, and computes composite scores with a confidence band drawn directly on the chart.

OPQ32

Behavioural personality — STEN 1..10

OPQ32

Verify G+

Cognitive ability — percentile

VerifyG

AMCAT

Technical skills — percentile

AMCAT

Composite radar

Upload one or more reports to populate the radar.

0 / 6 composites
No composites computable yet — upload at least one report.

Confidence

0–100

0.0

Weighted blend of coverage, reliability, recency, agreement.

Coverage0·35%
Reliability (α)0·25%
Recency0·20%
Agreement0·20%

Composites

Awaiting reports

Drop a PDF in any source on the left to populate the inference engine. The 6 composite definitions ship with the build; their scores appear here once inputs arrive.