Why is your tender team still overloaded?

The team is overloaded again.

More people would help. A better AI tool might help too. Something has to, because experienced commercial people are spending another week reading, checking, chasing and reconciling when they should be shaping the decision.

But what, exactly, is taking their time?

When bidder returns first arrive, the pressure comes from volume. Each response follows a different structure. Assumptions sit in different places. One bidder has answered the question directly, another has buried it in an appendix, and a third appears to have left it out altogether.

Later, there may be fewer bidders, but the work does not feel lighter. The preferred solution now touches technical, commercial, legal and delivery teams. Comments multiply, scope boundaries move and nobody is sure which appendix everybody reviewed.

Both situations look like the same capacity problem. They are not.

Capacity is the visible symptom. Underneath it, the commercial workflow is usually leaking time in one of two ways: the team cannot extract reliable signal from the information quickly enough, or it cannot coordinate the people and decisions that turn that signal into a commitment.

Until you know which leak is dominant, you risk buying help for the wrong problem.

Is information volume the real bottleneck?

Information volume is the bottleneck when capable people cannot reach a reliable comparison without first reading, sorting and restructuring every return. The work is not simply to understand each document. It is to turn several differently organised submissions into one decision-ready view without losing the evidence or the differences that matter.

You can see this when evaluators build their own matrices, search repeatedly for assumptions and exclusions, or discover late that two apparently similar answers were based on different scope boundaries.

An agent could map each return against the agreed structure, flag gaps and qualifications, and keep every observation linked to its source. A person still decides whether a difference is material, what clarification is fair and how it affects the award decision.

The outcome is not more summaries. It is a comparable, evidence-linked view that allows experienced people to start with the commercial question rather than the document hunt.

If that is where the hours are disappearing, extraction is the first workflow to redesign.

Or is the team losing control of the work?

Coordination is the bottleneck when the team understands the information but cannot keep the work aligned as it changes. The symptoms are familiar: unclear ownership, parallel versions, open interfaces and approvals that arrive after other people have already moved. The pressure comes from maintaining continuity, not reading faster.

A technical clarification changes a commercial assumption, but the contract appendix does not move with it. Legal reviews one version while delivery is planning from another. A senior lead becomes the only person who can explain the history, so progress depends on their memory and availability.

A useful system would make the current position, open interfaces, owners and approvals visible. Agents might connect changes or prepare the next action, but ordinary workflow rules may do part of the job better. The technology should follow the coordination problem.

People still negotiate, accept risk and approve commitments. The change is that they no longer have to reconstruct the state of the tender before they can exercise that judgement.

If this is where the pressure builds, coordination is the first workflow to redesign.

How do you choose the right problem first?

Choose the right problem by following one recent tender and finding where useful commercial work gave way to searching, comparing, chasing or reconciling. Do not begin with a platform shortlist or a catalogue of agents. Begin with the hours you would want your best people to spend differently next time.

Three questions are usually enough:

1. What did people have to find or prepare before they could make the decision?
2. Where did information, ownership or approval lose continuity?
3. If that work were prepared properly, what would experienced people do with the time instead?

The answer should point to one workflow, not an AI programme. It might be comparable evaluation, clarification control or preferred-bidder coordination. What matters is that the intervention changes a repeated commercial constraint rather than adding another tool around it.

This is the problem Commercial Discovery is designed to sort. It follows one real workflow, establishes whether the dominant leak is extraction, coordination or the handover between them, and identifies the first change worth designing before anybody commits to a build.

The result should be easy to explain: this is the work consuming capacity, this is what a system should prepare, and this is the judgement people will have more room to exercise.

Because your team does not need AI everywhere. It needs the right bottleneck removed.