Skip to main content
Workspace memory helps Custory AI carry useful context forward across the life of a workspace. Instead of every conversation starting from a blank page, the assistant can retrieve relevant prior knowledge from the same workspace and, when relevant, from the same journey.

​
What workspace memory is

Custory uses workspace-scoped memory powered by Supermemory. That means memory belongs to the workspace it came from. It is not one shared global pool across everything your team does. This matters because each workspace has its own:
  • Journeys
  • language
  • priorities
  • decisions
  • follow-up patterns

​
What memory helps with

Workspace memory is useful when the team wants AI to remember and reuse context such as:
  • Prior AI and chat context
  • Journey-related knowledge
  • Connected workspace context that has already been captured

​
Journey-biased retrieval

When the current AI conversation is tied to a journey, Custory can favor memory connected to that journey before falling back to broader workspace knowledge. That is important because it keeps the assistant closer to the local context of the work instead of surfacing unrelated information from elsewhere in the workspace.

​
Memory sources and ingestion

Supported workspace context can be synced into memory with metadata such as:
  • Workspace
  • Journey
  • Source type
  • Source ID
This creates more traceable retrieval than a vague “AI remembers something from before.”

​
Memory status tracking

Custory tracks memory source state rather than treating ingestion as invisible background magic. Memory sources can move through states such as:
  • Queued
  • Syncing
  • Ready
  • Failed
  • Deleted
This matters because teams need to know whether important context is available, unavailable, or intentionally removed.

​
Per-chat memory control

Custory includes a memory toggle in AI chat. Use memory on when:
  • You want the assistant to draw from prior workspace knowledge
  • The question depends on earlier journey or workspace context
  • You want less repetition across sessions
Use memory off when:
  • You want a narrow answer from only the current prompt
  • You are testing wording or reasoning in isolation
  • You do not want retrieval to influence the current chat
The setting is useful because not every AI task benefits from broader context. Turning memory off skips retrieval for that chat. Useful outcomes from the chat may still be learned as workspace context later.

​
When workspace memory is most valuable

​
Repeated strategic reviews

If the founder or PM reviews the same journey regularly, memory helps AI carry forward the team’s evolving understanding.

​
Ongoing product discovery

When the same themes return across chat, item edits, and follow-up work, memory helps reduce repetitive restating.

​
Cross-functional continuity

Memory becomes especially useful when support, product, and delivery are all contributing context over time.

​
Best practices

Memory is strongest when the underlying journeys and items are reasonably structured. Retrieval from noisy context produces noisier outputs, so better workspace hygiene leads to better AI help.
This is useful for focused drafting or when you want the assistant to reason only from what is visible right now. Use the toggle intentionally instead of leaving it on by habit.
Useful memory should help the assistant ground itself, but product judgment still belongs to the team. Retrieved context is a prompt aid, not a guarantee that the remembered statement is still true.

​
Memory habits that backfire

If the workspace context is vague or stale, memory will not magically make it precise. Clean up the source context first, then use retrieval to reduce repetition.
Sometimes a narrow prompt is better. Use the toggle intentionally so the assistant has the right amount of context for the job.
Memory helps the assistant remember. It does not mean every remembered statement is still correct, so keep validating important assumptions against the current workspace state.

​
Next step

Read AI as a workspace member for the broader AI collaboration model. Read MCP if your team also uses external AI clients.