No. There's no AI here, unless you bring your own. Connectors for common types; json-ld, RDF, whatever.

When someone asks "why did we decide this?" they're not doing archaeology through Slack threads.

They click the decision node. See the reasoning. Navigate to related decisions.

Time travel to past context. Without Sarah.

Every decision captures:
- What we decided
- What alternatives we considered
- What evidence we weighed
- What assumptions we're making
- What the next concrete action is

Not as a document to file away. As a graph you can search.

Like Git commits for knowledge work and workflows.

So I'm building something different: Assembly.

Not "better documentation" (documents die in Confluence graveyards).

Not "knowledge management" (nobody reads the wiki).

Infrastructure for institutional memory that doesn't depend on heroes.

Here's how it works:

Here's the thing about indispensable people:

Some WANT to be force multipliers. "I want my knowledge to scale, not be bottlenecked."

Some NEED to be gatekeepers. "My value comes from being the only one who knows."

Assembly is for the 1st group. It will threaten the 2nd group.

That's the point.

The pitch isn't "make everyone more productive."

It's "stop running institutional memory on hero employees who will eventually leave, burn out, or die."

Infrastructure for earned trust. Accountability through transparency. Decisions that don't depend on Sarah's brain.

If you try to "fix this" by documenting Sarah's work, she might resist.

Healthy orgs: "Finally I can share what's in my head."

Toxic orgs: "This threatens my value."

Either way - diagnostic information.

Assembly is for orgs ready to not force heroes to emerge. It helps heroes be more effective

So I'm building Assembly:

Infrastructure for institutional memory that doesn't depend on heroes.

Every decision is a graph node. Connected to evidence, actions, assumptions.

When someone asks "why did we decide this?" - they look at the graph. Not Sarah.

I've been Sarah. Not the hoarding kind, the drowning kind.

Institutional knowledge in my head because there's nowhere to put it.

Explaining the same context to 5 people because it's not captured anywhere.

Being human middleware is exhausting.

From leadership's view:
"Sarah gets so much done. Give her a raise."

From the org's view:
"If Sarah leaves, we're fucked."

From Sarah's view:
"My job security depends on being irreplaceable."

Stable dysfunction. Nobody can fix it without threatening Sarah.

Every company has a Sarah.

Sarah is the only one who knows how things connect.
Sarah has "the master spreadsheet."
Sarah works weekends.

Leadership thinks Sarah is a hero.

Sarah is actually a symptom of structural dysfunction.

Let me explain: 🧵