Important Decisions Were Happening, but Not Being Held

This note is about a condition I have seen more than once in organizational work.

At first, the problem does not usually announce itself as “decision architecture.”

It appears in more familiar operational language:

• meetings feel heavy,
• handoffs are unstable,
• the same topics keep returning,
• progress depends too much on specific people,
• and teams are never fully sure whether everyone is acting on the same understanding.

People often describe the situation like this:

• “knowledge is scattered,”
• “communication is messy,”
• “documentation is weak,”
• “too much depends on a few people,”
• or “we already talked about this, didn’t we?”

None of those descriptions is wrong.

But in many cases, they do not yet name the deepest layer of the problem.

What is actually happening is this:

Important decisions were happening, but the decisions themselves were not being held.

Judgment existed.
Choices were being made.
Work was moving.

But the decisions did not remain available in a form that supported:

• continuity,
• review,
• accountability,
• reuse,
• handoff,
• organizational memory,
• or scaling.

That is not only a communication problem.

It is a problem of decision-holding.

1. Decisions were happening constantly

The first thing to say clearly is that the organization was not lacking decisions.

Decisions were happening all the time.

They happened in:

• meetings,
• side conversations,
• Slack or chat threads,
• email exchanges,
• revisions to working documents,
• small verbal agreements,
• private notes,
• and the quiet habits of experienced practitioners.

So the problem was not absence of judgment.

The problem was that judgment was distributed across people, tools, and moments without becoming a stable organizational object.

This meant that teams could continue moving while still being unable to answer simple but operationally decisive questions:

• What exactly was decided?
• By whom?
• Under what assumptions?
• Against which alternatives?
• With what degree of confidence?
• Under what conditions should this be revisited?
• Who needed to know?
• And who actually knew?

Work continued.
But continuity weakened.

This is one reason teams often feel busy and active while still experiencing repeated uncertainty.

Movement exists.
But shared state does not.

2. A lot of the method existed only inside people

In one recurring type of case, the work had become heavily person-dependent.

That did not mean the people involved were doing anything wrong.

Usually it meant the opposite: they were competent enough to keep the system moving through experience, tacit judgment, and repeated local repair.

A skilled practitioner often had:

• a private sense of what “good enough” meant,
• a personal sequence for checking things,
• reference materials they kept close at hand,
• past examples they mentally compared against,
• and practical shortcuts for interpreting exceptions.

The problem was not that these resources were useless.

The problem was that they were rarely externalized in a form others could reliably use.

So the organization depended on knowledge that existed, but only in partial, private, or fragile forms.

This produced a common asymmetry:

the organization believed it had a process,
but in reality it often had people carrying a process internally.

That is a major difference.

A documented process can travel.
A person-carried process cannot travel very far without distortion.

3. Some of the most important assets were informal and local

A particularly important observation in these cases was that the real operating assets of the work were often not the official documents.

They were things like:

• a practitioner’s private checklist,
• example outputs stored locally,
• personal comparison materials,
• email threads used to answer recurring inquiries,
• remembered interpretations of past edge cases,
• or slides and notes that circulated only within a small circle.

These were not trivial artifacts.

They were often the materials people actually used when deciding.

But they remained:

• local rather than shared,
• practical rather than formalized,
• and useful only to those already close enough to the work to interpret them.

This is why I do not think the issue can be reduced to “we need more documentation.”

Many organizations already have documents.

What they lack is a way to turn operationally important local materials into shared decision infrastructure.

Without that move, organizations can look documented while still remaining structurally dependent on private memory and personal possession.

In other words, the organization may have traces of work while still lacking stable objects of judgment.

4. Email, chat, and meetings preserved traces, but not decision-state

Another recurring pattern was this:

the organization had many traces, but weak decision-state clarity.

For example:

• inquiry responses existed in someone’s mailbox,
• meeting discussion existed in calendars and notes,
• working assumptions appeared in draft slides,
• position changes appeared in chat,
• and revised directions appeared in someone’s spoken explanation.

So from one angle, the organization looked highly active and highly communicative.

But that did not mean it was holding decisions well.

Because preserving traces is not the same as preserving decision-state.

A decision-state includes more than “what was said.”

It also includes:

• whether the matter was exploratory or settled,
• who had authority,
• whether the conclusion was local or organizational,
• whether the position was final or provisional,
• what needed follow-up,
• and how the decision connected to other work.

When these layers are not explicit, people begin operating with different mental versions of the same organization.

One person thinks the direction is settled.
Another thinks it is still under discussion.
A third assumes it was already shared more widely.
A fourth believes someone else owns the final call.

From the outside, this looks like misalignment.

Underneath, it is often a failure of held decision-state.

That is why “we discussed it” is not the same as “we hold it.”

5. The issue was not only scattered information, but scattered authority

In some cases, the most damaging ambiguity was not informational but authoritative.

Teams did not only lack shared memory.

They lacked stable clarity on questions like:

• Who decides?
• Who recommends?
• Who reviews?
• Who escalates?
• Who informs others?
• Who can reopen the matter?

This matters because decision-holding is inseparable from authority design.

A choice cannot be held well if the organization cannot distinguish between:

• discussion,
• recommendation,
• draft position,
• working assumption,
• and final decision.

When these states collapse into each other, people leave the same interaction with different interpretations of what happened.

That produces a familiar cycle:

• a topic seems resolved,
• then reopens later,
• then partially regresses,
• then someone says “I thought we had already aligned on this,”
• and the team spends time reconstructing the state from memory.

This is not only frustrating.

It is structurally expensive.

It consumes time, weakens trust, and quietly increases dependency on the few people who still “remember what actually happened.”

6. Why decisions kept degrading over time

One of the most important practical observations was that poorly held decisions tend to degrade.

Not always dramatically.
Often quietly.

That degradation happens because the organization lacks a durable object that can resist drift.

Without a held decision object, what survives is often one of the following:

• a person’s memory,
• a fragment of an email thread,
• a note from one meeting,
• a slide that reflects one version of the story,
• or an assumption that gradually becomes “what we probably decided.”

This makes regression almost inevitable.

A team can move forward for weeks and still later find itself asking:

• Did we share this with everyone?
• Was this the actual direction, or only one proposal?
• Who agreed to this?
• Was this meant to apply everywhere, or only in this case?
• Did we ever formalize this, or just act as if we had?

This kind of degradation is not random.

It happens when a decision affects behavior but is not preserved as a stable reference.

The organization then begins to operate on partial recollection rather than governed continuity.

That is when old uncertainty returns disguised as new work.

7. Why this matters for scaling, outsourcing, and AI use

This condition becomes even more serious under pressure.

For example, when:

• more people join,
• work must be handed off,
• vendors need to participate,
• output must become more consistent,
• or AI is introduced into the workflow.

In all of these situations, weakly held decisions become a bottleneck.

Why?

Because scale requires portability.

And portability requires that judgment survive movement.

If decisions remain trapped inside:

• one inbox,
• one practitioner,
• one meeting room,
• one draft document,
• or one private interpretation,

then the workflow cannot scale cleanly.

The same applies to AI-era operations.

AI can support drafting, summarization, and analysis.

But if the organization cannot clearly hold:

• what has been decided,
• what remains open,
• what counts as authoritative,
• and which conditions allow revision,

then AI operates inside ambiguity rather than helping resolve it.

In that case, AI may speed up content production while leaving decision continuity weak.

That makes the system faster, but not more governable.

This is why AI often reveals structural weakness that already existed.

It does not create the ambiguity.
It exposes how little of the organization’s judgment was actually holdable.

8. What changed when the problem was reframed

A useful shift happened when the issue was no longer treated as “communication mess” alone.

Once it was reframed as a problem of decision-holding, different questions became possible.

For example:

• Which decisions actually need explicit holding?
• What minimum structure should a held decision contain?
• Where should that decision live?
• How should related discussions link to it?
• What counts as a recommendation versus a decision?
• Who owns review?
• Which supporting materials should become common assets rather than personal tools?
• How should email-based judgment be converted into reusable organizational memory?
• How should decision-state be transmitted across teams and roles?

This changed the design goal.

The goal was no longer “document more.”

It became:

hold the judgments that shape coordination in a form the organization can continue from.

That is a different ambition.

And a more useful one.

Because otherwise teams usually swing between two weak options:

• over-documentation no one uses,
• or under-structured communication no one can reliably reuse.

Decision-holding is the middle path between them.

9. What this taught me

This kind of case taught me that good judgment in the moment is not enough.

A decision only becomes a true organizational resource when it can remain available afterward.

That means the quality of decision-making has at least two dimensions:

1. Was the judgment sound?
2. Was the judgment held in a way the system can continue from?

The second question is often neglected.

But in practice, it changes everything.

A strong decision that disappears into memory is weaker than it first appears.

A decent decision that can be reviewed, reused, revised, and linked to future work may be far more valuable.

This is why I increasingly think that organizational memory should not be treated as passive storage.

It should be treated as an active part of decision architecture.

Not a warehouse of leftovers, but a structure that keeps judgment alive across time.

Closing

Important decisions can be happening every day while still failing to become part of the organization’s usable memory.

This is one of the most common reasons teams feel busy but structurally unstable.

The work moves.
People adapt.
Messages accumulate.
Documents grow.

But continuity remains expensive.

Review stays difficult.
Authority remains blurry.
Knowledge does not compound.
And the same judgment work returns again in slightly different forms.

That is why I think decision-holding deserves to be treated as a real design problem.

Not as an afterthought to communication.
Not as a matter of note-taking hygiene.
And not only as documentation.

But as part of how organizations make judgment portable, reviewable, and survivable across time.

If judgment matters, then how judgment is held matters too.