Why do unclear tenders produce incomparable bids?

The tender closes on time. Every supplier has followed the requested structure. The evaluation team should now be able to compare like with like.

Then the questions begin.

One supplier priced an interface the buyer thought was included elsewhere. Another treated an output as optional. A third proposed a different delivery basis because the acceptance evidence was not clear. Each response may be credible, but the evaluators are no longer comparing the same commercial proposition.

The immediate temptation is to blame the suppliers or add another scoring workshop. Often, however, the variation began inside the tender.

A tender does more than request offers. It creates the basis on which suppliers make assumptions, allocate risk and decide what evidence to provide.

Why can a compliant tender still produce incomparable bids?

A compliant response is not necessarily a comparable response. If the requirement, interfaces, evidence or evaluation method leave material room for interpretation, suppliers will fill that space differently. The tender may generate several valid answers to several subtly different questions.

That matters because the differences do not stay inside the technical response. They can appear in price, programme, exclusions, dependencies and risk allowances. The evaluation team then spends time reconstructing each supplier's basis before it can judge the offer itself.

For covered UK public procurements, current Cabinet Office guidance says award criteria must be clear, measurable, specific and proportionate, and accompanied by an assessment methodology. Those are legal requirements in that context. They also express a useful commercial design principle for any tender: suppliers need to understand what decision their evidence must support.

More explanation does not always solve the problem. A long pack can still be unclear if its documents disagree, interfaces have no owner or the evaluation criteria reward something different from the stated outcome.

What should be settled before the tender is issued?

The buyer does not need every future detail to be known. It does need a controlled basis for competition. Before issue, five elements should tell one coherent story:

• The decision: what the organisation is selecting and which outcomes matter.
• The requirement: what is mandatory, where flexibility is welcome and how alternatives should be presented.
• The interfaces: what sits with the supplier, the buyer or another party.
• The evidence: what will demonstrate capability, quality, deliverability and commercial credibility.
• The evaluation: how the evidence will be assessed and how material differences will affect the decision.

This is not simply document quality. It is alignment between commercial strategy, technical requirements, risk allocation and evaluation.

A useful pre-issue review therefore asks more than whether every section is present. It asks whether two capable suppliers could read the same material and reasonably build offers on different commercial bases. If they could, the tender is not yet doing enough work for the evaluation team.

Can AI help prepare a clearer tender?

AI can help with the exhaustive preparation, but it should not decide the commercial position. A governed agentic workflow could compare documents, identify inconsistent terms, map requirements to evaluation criteria, trace defined interfaces and flag where an evidence request has no corresponding decision use.

It could also prepare questions for human review:

• Where does the pack use two definitions for the same obligation?
• Which requirement has no owner or acceptance evidence?
• Which evaluation criterion cannot be traced to the stated outcome?
• Where could a supplier make a material assumption without declaring it?

The output would still need subject-matter, procurement, legal and commercial review. The agent has not chosen the scope, accepted the risk or designed the award decision. It has made contradictions and gaps easier for the accountable people to see before suppliers have to price them.

That distinction matters. The opportunity is not an autonomous tender writer. It is a controlled system that performs repeated cross-checking while the team applies the judgement that gives the tender its meaning.

How do you know the tender is ready to issue?

A tender is ready when the organisation can explain the same commercial basis from four viewpoints: the outcome it needs, the work suppliers must price, the evidence it will assess and the risks each party will carry.

Before release, ask:

1. Could suppliers identify every material assumption they need to make?
2. Do the requirements, contract, pricing structure and evaluation method agree?
3. Can evaluators explain what evidence would distinguish a stronger response?
4. Is every important interface owned?
5. If a clarification changes the basis, can the team identify every document and supplier affected?

If those answers are weak, issuing faster will not create a cleaner competition. It will move the unresolved work into clarifications, qualifications and evaluation.

The quality of the offers begins with the quality of the question the market receives.

Sources and scope

• UK Cabinet Office, Guidance: Assessing Competitive Tenders, updated 17 August 2026: https://www.gov.uk/government/publications/procurement-act-2023-guidance-documents-procure-phase/assessing-competitive-tenders-html
• UK Government Commercial Function, Module 7: Assessment and award of contracts: https://www.gov.uk/government/publications/the-official-procurement-act-2023-e-learning/module-7-assessment-and-award-of-contracts
• The public-procurement requirements are not presented as legal advice or assumed to govern private-sector tenders. The wider application is Ignis design guidance.