Organizational Memory

  1. 01Intake
  2. 02Requirement
  3. 03Design
  4. 04Work order
  5. 05Build
  6. 06Testing
  7. 07Sign-off
  8. 08Memory

The record of what was decided and delivered on this estate. After a change ships, the design, the evidence, and the named approvals stay in the workspace. The next request starts from that record, not from a blank page.

In the workspace

One workspace. Every part of delivery.

Command Center is the map of this estate. Decisions, evidence, and named approvals stay here so the next request starts from the record.

What it holds

The workspace keeps the change.

Every change to an application stays in the same workspace. The technical design versions forward. Each new requirement adds tasks on top of the ones already there, so the next person does not re-learn the estate by hand.

Captured

Decisions, exceptions, outcomes

Requirements, designs, approvals, evidence, and the blast radius of what shipped. Human judgments at the two gates are part of the record, not a comment in a ticket.

Used for

The next request

Gate 8 feeds gate 1. Subsequent work inherits the environment, the standards already encoded, and what the last pass learned.

How it is stored

Four tiers. One customer boundary.

Retrieval is scoped by workspace and permission. It never crosses a tenant. Customer content does not move between customers. Patterns learned on one engagement become candidate Genes on the next only after review, and only as abstracted patterns.

TierHolds
HotActive workspace state and in-flight task context.
WarmRecent engagement history and current instance metadata.
ColdHistorical engagements and archived evidence.
GraphDependencies, relationships, and learned patterns used for impact and blast-radius traversal.

Already encoded

500+ skills. 152 specialist agents.

A Gene is a reusable encoded standard: a policy, a delivery pattern, a control, a testing rule, a best practice, or a decision standard. Encode it once. It runs on later work without being taught again.

The platform ships with 500+ ServiceNow skills already encoded, and 152 specialist agents in five functional groups. Those two figures are structural counts, not modelled outcomes.

Ask Gene answers from the instance and the workspace memory, and can route work, under the same two gates as every other surface.

Loop

The next request starts here.

Memory is the start of the next factory pass, not an archive at the end.