When you work with the Pecan agent, it can remember useful facts about your data from one conversation to the next - so you don't have to re-explain the same things every time. This happens automatically in the background; there's nothing you need to set up.
How It Works
As you chat with the agent - for example, explaining what a column means, how tables relate to each other, or a quirk in your data - the agent picks up on the facts you share and stores them.
The next time you're working in a flow, the agent can draw on what it already knows about your data. When it does, it will tell you, using phrasing like "based on what I know about your data" or "from previous conversations" - so it's always clear when the agent is using something you told it before versus something from the current session.
Memory is shared across your organization's flows - a fact you share in one flow can help the agent in another, not just the one you're currently working in.
Viewing and Managing Memories
You can see everything the agent has learned, and manage it, from the Memory tab at the top of the platform.
Each entry shows the fact that was captured, which flow it came from, who it came from, and when it was captured.
From this screen you can:
Search - find a specific memory using the search bar.
Sort - order the list by date or by Flow ID.
Filter - narrow the list down to memories from a specific user.
Delete - remove a memory using the trash icon. This is useful if something was captured that's no longer accurate or that you'd rather the agent not use going forward.
Good to Know
Only things stated by you or your teammates are remembered - the agent doesn't store its own suggestions or guesses as memory.
There may be a short delay between saying something in a conversation and seeing it appear in the Memory tab.
Deleting a memory stops the agent from using it in future conversations - it doesn't change or undo anything already discussed.