Buyer-side tender evaluation assurance

Turn tender scoring into a defensible award record.

When the award is questioned, the answer is already written. Blind scoring, sealed prices, code-composed marks and a challenge-ready audit trail — one controlled workflow for any team that has to show how it chose. Public sector, housing, utilities, universities and private-sector procurement running a formal tender.

See a live demo Open the app

UK data residency Cyber Essentials certified Procurement Act 2023 ready In-VPC / zero-egress option

Before the solution, the problem

A contested award unravels in one of three ways

Award IQ is built to close all three — before the challenge lands.

“The panel’s scores are all over the place — and you can’t explain the spread.”

Code composes every mark from the same enforced rubric, so scores don’t wander or anchor at “a comfortable 4.”

How scoring is controlled

“A price slipped into the quality scoring — and the process is now challengeable.”

Two-envelope seal freezes quality before any price is revealed. Prices structurally cannot touch the scores.

See two-envelope integrity

“The award gets challenged — and you’re reconstructing the reasoning from email chains.”

An append-only audit trail and a ready-written Challenge Defence Pack mean the answer exists before the question does.

See the defence pack

Why it wins

Three things that close the award

Consistency

Code composes the score, so marks don’t wander or anchor at “a comfortable 4”. Every response is judged by the same enforced rules.

Impartiality

Genuinely blind quality scoring, sealed before price, with a clean two-envelope reveal. Bias designed out, not asked for.

Defensibility

Every score has a traceable reason, every stat is code-verified, and the whole process is enforced and recorded — the challenge answer is written before it’s asked.

Complete capability

Everything between bids in and award signed off

The layer the mass-market portals are thin on — score, evidence, moderate, stress-test and defend.

Evaluation workflow

One controlled, repeatable process the whole panel follows — nothing skipped, nothing ad-hoc.

Guided multi-step flow

Setup → upload → score → seal → price → reports → approve → lock. The same disciplined path every time.

Auto-extraction of questions, weightings & word limits from the ITT

The framework is set up in minutes — no re-keying.

Auto-extraction of per-question bidder responses (DOCX/PDF)

Submissions mapped to questions automatically, ready to score.

Past-evaluations drawer with instant load

Re-open any evaluation — scores and every report appear immediately, no regeneration.

Document classification (Procurement Sensitive / Official)

The right sensitivity marking stamped on every export.

Scoring engine — controlled, not a black box

The engine analyses. Code sets the final mark. Consistent, reproducible, defensible.

3-stage pipeline: decompose → dimension-score → code-composed 0–5

The engine reasons; code composes the score — marks don’t wander or anchor at “a comfortable 4”.

Four rubric dimensions — Evidence · Understanding · Methodology · Coverage

Every score shows why: you see exactly which lever drove a 3 vs a 4.

Matrix caps enforced in code (missed requirements capped; non-response → 0)

Your published scoring rules applied identically to every response.

Calibration corpus that learns from your overrides

The engine drifts toward how your organisation actually scores.

Provenance stamp on every score (engine version + models + time)

At challenge you can prove exactly how each mark was produced.

Empty-response guard

Only a genuinely blank answer scores 0 — no bidder is unfairly zeroed.

Evaluator review & override

Human judgement always wins — and nothing changes silently.

Per-bidder review screen

Response + rationale + strengths/weaknesses + evidence, question by question, in one readable place.

Inline re-score (0–5) with a required reason

Evaluators can override — but must say why.

Change history per question (who, what, when, why)

A visible, per-question trail of every re-score.

Cast-iron identity

Every change is attributed to the verified logged-in user — the real person, not a dropdown pick.

Auto read-only once sealed / locked

Overrides can’t happen after the envelope closes — integrity preserved.

Impartiality — true blind scoring

Genuinely blind out of the box, with deliberate, auditable un-blinding when you choose.

Blind mode on by default

Reference codes, not names — fair, unbiased scoring from the start.

Content redaction

Masks names and aliases inside the response text and the generated narrative — genuinely blind, not just the labels.

One-click Reveal names (action-labelled)

Deliberate, auditable un-blinding on your terms.

Two-envelope integrity

Prices cannot influence quality — the core probity control, enforced.

Seal quality before pricing

The moment quality is sealed, prices can’t touch the scores.

Sealing freezes quality scores

No quiet edits once prices are in play.

Reversible un-seal (audit-logged)

Correct a genuine quality error without losing the defensible record.

Commercial envelope + Calculate & reveal winner

Enter costs against the right bidder, then reveal the award.

Price & MAT

The award decided on the correct basis — quality and price — with the maths spreadsheets can’t do.

Three selectable UK price formulas

Lowest÷bid · linear · max–min — match whatever your ITT published.

Combined MAT ranking

The Most Advantageous Tender, computed and ranked.

Exact £ price-sensitivity

“X would need to cut its £120k bid to £85k to win.” The one question a spreadsheet can’t answer — award-defence maths.

“Quality-decisive even at £0” detection

The strongest possible line when price couldn’t have changed the result.

Reports & governance documents

The whole evidenced pack — written for you, with every stat verified by code.

Moderation Report

The signed evaluation record: per-question rationale + consistency flags.

Challenge Defence Pack

A ready-made rebuttal defending why each bidder did or didn’t win.

Assessment Summaries (per-tenderer)

“How you were assessed, and why the winner won.” Procurement Act-shaped and standstill-ready where that applies; a straight, defensible debrief where it doesn’t.

Bidder Feedback Letters (printable / PDF)

Clear debrief letters, one per bidder.

Code-stamped names & every stat

The engine writes only narrative — it cannot misname a bidder or invent a figure.

Citation verification

Fabricated quotes are flagged — the evidence trail is real, not hallucinated.

Full-pack export + “what changed” diff + completeness guard

One button hands over the whole classification-marked pack; a report that missed a bidder is flagged, never silently short.

Defensibility & governance — the moat

When the award is questioned — by an unsuccessful bidder, internal audit or your own board — the answer is already written.

Append-only audit log (server-stamped actor + IP + time)

An un-spoofable record of who did what, when.

Human-approval state machine

Draft → reviewed → moderated → approved → locked. You can’t lock until it’s approved.

Conflict-of-interest declarations enforced before approval

No award proceeds until every panellist has declared.

Lock & sign-off (immutable + typed attestation)

A frozen, signed record for the file.

Platform, security & support

Built for both SaaS and the most sensitive in-house procurements.

Hard multi-tenant isolation

Structural — no un-scoped DB access. One organisation physically cannot see another’s data.

Auth seam (Clerk cloud / OIDC-Cognito VPC)

Fits both SaaS and enterprise identity.

In-VPC / Amazon Bedrock option

Zero data egress over PrivateLink — deployable inside your own environment.

Iris — agentic in-app support

Inspects the real account for fast, accurate help without a support queue.

Tested award maths (39 unit tests)

Trust in the numbers — failures surface instead of hiding silently.

Where it sits

Keep your portal. Use Award IQ to win the challenge.

Not a full source-to-pay suite — deliberately. Award IQ owns the layer the mass-market portals are thin on: after bids are submitted, before award is signed off. Score it, evidence it, moderate it, stress-test the price, and defend the award — all in one controlled, recorded workflow.

ScoreEvidenceModerateStress-testDefend

Make your next award challenge-proof.

See Award IQ score a real tender pack end-to-end — blind, sealed, evidenced and defended.

See a live demo Open the app