Why do tenders inherit decisions without their history?

I have sat in enough tender kick-offs to recognise the moment.

Everyone arrives ready to move. The opportunity is exciting, the deadline is real and capable people are already thinking about the work ahead.

Then someone asks a simple question.

Why did we choose this route?

The room hesitates, not because nobody knows, but because the answer is scattered. Part of it sits in an early model, part came from a conversation months ago, and the person who understood the concern best is not in the room.

Of course the team will find the answer, but by the time it does, finding it has already become part of the tender.

Months later, the same unease returns when a commercial assumption is sitting inside a price and finance, legal or delivery wants to know how the business arrived there.

And there is a cost to it. Good people feel exposed, senior leaders lose confidence in what is in front of them, and the team becomes defensive just when it most needs to think clearly.

Nobody has failed, but the business has lost part of its commercial memory.

Why is the tender only the middle?

The tender inherits decisions made before it and turns them into commitments people will live with afterwards. It is not an isolated island, but a bridge between the opportunity people imagined and the deal the organisation may eventually sign.

That bridge needs more than documents. The team needs to know:

• what the organisation believed;
• what evidence supported it;
• what has changed; and
• who can make the next commitment.

Without that memory, a provisional route starts to sound settled, an optimistic date slips into the plan, and a concern meant to be revisited survives into the offer.

What does weak commercial memory cost?

Weak commercial memory costs more than time spent searching for the latest file. You can feel it when a confident discussion becomes cautious and somebody says, “I think that was agreed,” only for someone else to ask, “Agreed by whom?”

Then the cost compounds:

• settled questions reopen;
• teams reconcile competing versions instead of making the next decision;
• assumptions travel further than the evidence behind them; and
• the person who remembers becomes the unofficial thread holding the opportunity together.

That works until they are unavailable, stretched or trying to recall a decision made months earlier under different conditions.

Commercial excellence is not a perfect process map. It is the ability to keep good judgement alive as the opportunity moves, so people can challenge a position without first having to reconstruct it.

Why does agentic AI make this practical now?

This is where I am genuinely excited by agentic AI. In 2026, we can begin connecting enterprise knowledge, working context and tools around a commercial decision while keeping authority with people.

Imagine a commercial review that does not begin with ten minutes of asking which file is current. The original assumption, its evidence and the outstanding decision are already visible, leaving the room to focus on what requires experience and courage.

Or imagine a clarification changing an assumption agreed months earlier. Instead of relying on the one person who remembers, the team can see where the assumption entered the solution, what else it touched and who now needs to decide.

By final approval, the numbers may look settled even though the evidence beneath them has moved. An agent can bring that history and the unresolved conditions into the room before a leader commits.

But this only earns trust when people can see where an answer came from, where uncertainty remains and who must make the call. Otherwise, AI does not repair commercial memory. It hides the break behind a confident answer.

Where should a commercial leader start?

Start where commercial memory already fails: the review that always overruns, the assumption everyone handles carefully or the handover followed by private messages.

Choose one recent opportunity and ask:

1. What did the next team have to rediscover?
2. Which assumption changed without a clear reason or owner?
3. What could a system prepare while a named person still makes the decision?

The answer may be a better decision record, a clearer ownership rule or one bounded human-agent workflow. What matters is that the organisation stops asking its best people to carry the whole story in their heads.

Every hour spent looking backwards is an hour the team cannot spend shaping what comes next.