Tender Intelligence QA

Tender Intelligence QA is a dedicated quality review that runs across your entire tender response. It analyses your answers for issues that are hard to spot manually: weak coverage, tonal inconsistencies, scoring criteria missed, and gaps in evidence: and surfaces them as actionable findings.

When to use it

• After bulk generation: to quickly identify which questions need the most attention.
• During refinement: to check whether your edits have addressed the issues from a previous audit.
• Before submission: as a final quality gate to catch anything you may have missed.
• Run QA at least once before export; ideally run it twice (once after generation, once after your editing pass).

How to launch QA

• Open the project page for the tender response.
• From the Status tab in the right-hand sidebar, click Run New Audit, or look for a Tender Intelligence QA button in the top actions area.
• The audit runs as a background job. A progress indicator will show. For large responses this may take a couple of minutes.
• You will receive a notification when the audit is complete.

Reading the report

• Open the full report using Open full report. Findings are grouped as Whole response and By question, so issues that apply to the bid as a whole are kept separate from issues tied to a single answer.
• Whole response findings — for example a strategic weakness or an inconsistency that spans several answers — do not have a single question to attach to. They surface here and in the Tender Intelligence QA widget on the project page.
• Each finding includes a category and a description of the issue.
• Answers with uploaded photos or embedded diagrams receive an additional visual review. Findings may flag missing captions, sizing issues, or diagrams that do not match the answer content.
• Severity indicators help you prioritise: address high-severity findings first.
• A Readiness banner gives you a summary read on how close the response is to being submission-ready, based on how many issues remain open.

Fixing issues automatically

• After a run finds issues, you can choose Yes, fix what you can. Uplift works through the open issues one at a time and applies fixes directly to the affected answers.
• Once it has worked through the issues, a fresh check runs automatically to confirm what was resolved. Anything that could not be fixed automatically stays open for you to review.
• This does not replace your own judgement — always read through what changed before treating a question as finished.

Acting on findings manually

• Where findings link directly to a question, click through to open that question and make changes.
• For tonal issues, use Eva to rewrite sections of the answer, guiding it with your Standing Instructions and preferred tone.
• For coverage gaps, check whether the Content Vault has relevant material: if not, add it and regenerate.
• For scoring criteria missed, look at the question text carefully and ensure your answer addresses every point the buyer is looking for.
• Mark a finding resolved once you're satisfied it no longer applies — this updates the open-issue counts shown across the project.

Past-proposal feedback

• If your organisation has captured feedback on past proposals (via Document Context in the Content Vault), that feedback is included in the project-level QA analysis.
• The audit will flag it if a current answer repeats a mistake identified in previous evaluator debriefs or client feedback.
• This check only applies when relevant feedback exists and only where a lesson genuinely applies to a specific answer — it will not invent issues.

Re-running QA

• After making edits to several questions, run a fresh audit to see whether the issues have been resolved.
• Each audit run is independent: previous reports are stored so you can compare before and after.
• A clean second audit (with no high-severity findings) is a good signal that the response is ready for export.