AI will expose your slowest bid decision.

The tender has been analysed. The first response is taking shape. The gaps are visible.

Then the bid stops.

It needs a decision on whether the opportunity is still worth pursuing. Or which solution the business is prepared to stand behind. Or whether a contractual risk is acceptable.

The team has moved faster. The decision has not.

This is the constraint AI is about to expose in many bid organisations.

On 6 September, OpenAI published data on how its own researchers are using coding agents. Researchers are contributing code faster, running more experiments and delegating more complex work. But OpenAI also reported that complex tasks still require significant human steering.

Its most useful observation for business leaders was not about the technology itself. As more work is automated, the least automatable tasks take a larger share of human effort and become the new bottlenecks.

This is evidence from OpenAI about work inside OpenAI. It is not evidence about bid performance. But the operating pattern should feel familiar to anyone responsible for complex pursuits.

AI can accelerate requirements analysis, research, evidence retrieval, drafting and checking. Yet a bid can still wait days for someone to decide:

• Do we know enough about the customer to bid?
• Which solution are we prepared to commit to?
• Is the evidence strong enough to support this claim?
• Will we accept this commercial or delivery risk?
• Is the final submission ready to carry our name?

These are not production tasks. They are decisions with consequences.

Why can faster bid production create a longer queue?

Most early AI business cases measure task speed: how quickly tender analysis, drafting or checking can be completed. Those measures are useful, but they do not show whether the pursuit itself moves faster.

If more work reaches the same small group of decision-makers, faster production simply creates a longer approval queue. Commercial, technical and executive leaders receive more choices or exceptions to resolve, often without the context they need to decide confidently.

Production capacity only becomes commercial capacity when the organisation can make the decisions that release it.

How should leaders measure bid decision delays?

Before investing in another AI use case, take one completed bid and reconstruct its journey from opportunity to submission.

Separate the time into two categories:

1. Time spent producing or improving the work.
2. Time spent waiting for a consequential decision.

For each important gate, ask five questions:

• What decision was required?
• When did the need for that decision become visible?
• Who had the authority to make it?
• When was it actually resolved?
• What work stopped, continued at risk or had to be redone while the team waited?

The aim is not to blame a busy executive for taking too long. Decision latency is often a system problem.

The information may arrive in the wrong format. The owner may be unclear. The decision may be brought forward before the evidence is ready, or far too late for the answer to change the bid. Several people may believe they have approval rights, while nobody feels personally accountable for giving the answer.

That is why adding another reminder or meeting rarely fixes it.

What should AI prepare for a bid decision?

An AI-supported workflow could assemble the context, trace the supporting evidence, identify unresolved assumptions and prepare options for review. It could show what changed since the previous gate and flag when an agreed response time is about to be missed.

That would make the human decision easier to reach and harder to avoid.

But it should not silently decide whether to pursue the opportunity, accept the risk, commit the solution or release the submission. Those decisions belong to named people with the authority and accountability to make them.

The goal is not to remove senior judgement from the bid.

It is to stop wasting that judgement on reconstructing context, searching for the latest version or discovering too late that everyone was waiting for somebody else.

Which bid decision is constraining your organisation?

As AI makes more of the production layer faster, organisations will get a clearer view of what has really been slowing their bids down.

In some businesses, it will be customer insight. In others, solution alignment, commercial assurance or final approval. We do not yet have enough bid-specific evidence to say which constraint dominates.

But every organisation can identify its own.

Do not begin by asking where AI could save the most hours.

Begin by asking which important bid decision regularly arrives too late, who owns it and what that delay costs the pursuit.

Because once production speeds up, your slowest decision becomes the speed of the bid.