Rovo has been quietly getting to know you for a while. Every issue you touch, every page you write, every PR you review flows into the Teamwork Graph. Rovo uses that to personalize what it shows you and how it helps.
Up to now, that’s all been behind a one-way mirror. Rovo could see in. You couldn’t. You couldn’t see what it had learned, correct it when it was wrong, or say, “That Aurora project? I shipped it in Q1, remember to stop surfacing it in every summary.”
That changes today. Profile Memory in Rovo Chat is now generally available, giving every user:
-
A clear view of what Rovo knows about them
-
Simple controls to shape, correct, or delete that memory
I’m Manav Trivedi, Senior Product Manager on Rovo Chat. Here’s how Rovo Memory works and how to get real value from it.
🧠 Two kinds of memory
Rovo Memory is built from two sources:
-
Implicit memory
-
What Rovo infers from your Teamwork Graph activity
-
Includes your role, projects, collaborators, and the work you keep coming back to
-
Powers “day-one” personalization, even before you’ve told Rovo anything directly
-
Explicit memory
-
The instructions you deliberately give Rovo
-
Examples: “Always use Australian English.” “Never use emojis.” “Always add a TL;DR to every page you create.”
-
Say it once and it persists across chats and sessions
Together, they make Rovo feel less like “an AI assistant” and more like your assistant.
🎛️ The controls you now have
Open Rovo Chat → … menu → Settings and Memory to:
-
See it
-
Read the profile Rovo has built: how it describes your work, projects, collaborators, and the saved items (docs, tickets, links) it’s holding onto.
-
Suggest edits to your summary
-
Propose changes to your profile summary (e.g., remove an old project, update your focus area).
-
You’ll see your summary updated within 24 hours.
-
Delete explicit instructions
-
Remove past instructions you no longer want Rovo to follow.
-
Once deleted, Rovo stops using that explicit preference.
-
Toggle chat-derived memory
You stay in control: you can see what exists, refine it, or wipe it.
💡 A concrete example
Say you write a weekly status update every Friday.
-
The first few times, you tell Rovo what you want:
-
Bullet points
-
Grouped by project
-
Blockers at the top
-
No fluffy intro
-
Then you say: “Remember this as my preferred weekly update format.”
From that point on, every Friday you can just say “Draft my weekly update” and Rovo:
Change teams or change your format? Open Memory, update the preference, and Rovo adjusts.
⚖️ Implicit vs Explicit vs Your Controls
|
|
What it does
|
Best for
|
Example
|
|---|
|
Implicit
|
Learns automatically from your Teamwork Graph activity
|
Day‑one usefulness, cold starts, ambient personalization
|
“Rovo already knows I’m a PM focused on Rovo Chat.”
|
|
Explicit
|
Durable rules you set yourself
|
Format preferences, tone, hard rules
|
“Always lead with a TL;DR. Never use emojis.”
|
|
Memory
|
See / suggest edits / delete / toggle everything
|
Trust, correctness, privacy
|
Delete: “Aurora project, shipped in Q1.”
|
The combination is what makes Rovo both smart and correctable.
🧹 A few housekeeping details
-
Per user, per site
-
First‑run experience
Rovo Memory is what turns Rovo from “an AI assistant” into your assistant:
-
It knows what’s happening across your work via the Teamwork Graph
-
It knows how you like to work because you told it
-
And now, you can see all of that, fix any of it, and delete what you don’t want — whenever you want
Try it out and tell us in the comments 👇
What did Rovo get right about you? What was the first memory you saved or deleted?
