Skip to main content

Departments

The same thing goes wrong in every team.

A renewal, a hire, a supplier, a contract clause. Different words, same failure: the decision survives and the thinking behind it does not. Nine teams record decisions in Clyreon, and they all share one memory rather than getting a tool each.

What goes wrong, and what is left behind

Nine functions. The questions are examples, not customers.

Customer Success

Do we renew this account at current terms, or offer a concession?

The discount gets given, the account renews, and a year later nobody can say whether the concession bought loyalty or simply cost margin.

Kept instead: The reason for the concession, the evidence it rested on, and whether the thing you were worried about actually happened.

Sales

Do we pursue the expansion this quarter, or protect the base renewal first?

A deal is called at 80 per cent and slips twice. The forecast is corrected each time, so the original call disappears and nobody learns whose calls to trust.

Kept instead: What each person predicted and how sure they were, checked against what closed.

Finance

Do we fund the plan at the requested amount, a reduced one, or defer it?

Next year the same team asks for the same budget, and the argument starts from scratch because the conditions attached to last year's approval were in an email.

Kept instead: What the money was approved on the strength of, and whether that turned out to be true.

Product

Do we prioritise this ahead of the committed roadmap, and what are we displacing?

Six months on, the feature has not moved the number it was meant to move, and the case for building it has been overwritten by the case for the next thing.

Kept instead: What you expected the feature to do, what you gave up to build it, and what actually happened.

Procurement

Do we renew this vendor, re-tender, or extend on current terms?

The contract auto-renews because the person who ran the last review has left and nothing they knew went with the file.

Kept instead: Why this supplier was chosen last time, what the alternatives were, and what the concerns were.

Vendor Risk

Do we accept this vendor's security posture, or require remediation first?

An exception is granted for a quarter. Two years later it is still in force and nobody remembers it was ever meant to be temporary.

Kept instead: The exception, the condition attached to it, and an alert when that condition comes due.

Legal

Do we accept the exposure as drafted, redline it, or escalate to outside counsel?

The same clause is negotiated differently by three people in one year, because the reasoning behind the first position was never written anywhere findable.

Kept instead: The position taken on that clause, why, and every other contract that cited the same reasoning.

People

Do we hire this candidate, hire at a lower level, or keep searching?

The hire works out badly and the debrief becomes a search for who to blame, because nobody wrote down what the doubts were beforehand.

Kept instead: The doubts that were raised at the time, by name, and whether they were the ones that mattered.

Support

Do we escalate this account to executive attention, or resolve and monitor?

The account churns. The post-mortem finds three warnings, and an argument about whether anyone could reasonably have acted on them.

Kept instead: What was known on the day the call was made, so the question is answerable rather than arguable.

All nine ship today. Adding a tenth is a name and some wording. Nothing changes underneath, and no existing decision is touched.

What actually changes between teams

Only the wording. That is deliberate, not a simplification.

A team in Clyreon is a name, the kind of decision it makes, and the words on its screens: what it calls the thing being decided, how it labels the deadline and what is at stake, the positions it offers, and the kinds of evidence it expects.

That is all of it. A team stores no data of its own and holds nothing back from the rest of the company. If it did, the promise that nine teams share one memory would not be true. The suggested positions are suggestions only. The moment a system tells Finance which words it may use, it has stopped being a record of that company’s judgment.

Choose a department
What is being decided

This team calls it an account.

Deadline

Labelled Renewal date, and it starts warning 30 days out.

Wording only. Clyreon itself has no notion of late. A renewal needs a month of runway to negotiate. An incident needs hours.

Stakes

Labelled Annual value

Positions

Renew · Renew with concession · Do not renew · Escalate · Defer

Suggested by the interface, enforced by nothing.

Evidence

api-poll · webhook · meeting-note · survey · manual-entry

Usually owned by

Customer Success, Account Management

Everything not listed above is identical across all nine: the decision itself, the evidence, the reasoning, how sure people were, what would prove them wrong, who disagreed, the outcome, and the access log.

Why one shared memory, and not nine tools

Split the memory up by team and three things stop working.

Evidence crosses

A fact is stored once and can be used by any number of decisions. The same note about a customer’s reorganisation supports the renewal decision, the support escalation and next year’s forecast. One fact used three times, not three copies drifting apart.

Consequences cross

When a fact turns out to be false, Clyreon can name every decision now worth another look, including ones in other teams that never used the fact and only relied on a decision that did. A memory split up by team cannot follow a false claim across that boundary.

Precedent crosses

A procurement decision and a security review of the same supplier are two separate decisions, and each one links to the other. That link is what lets somebody, two years later, read one and find the other.

Teams in Clyreon separate what people see and how they work. They never separate the memory itself, and they were never able to.