Phoebe Product Update

Memories, follow-up queries, and the next evolution of Phoebe’s subagent architecture — here’s what we shipped this month.

🧠 Memories

Phoebe now lets you create and manage Memories — a way to give you direct control over how Phoebe behaves and learns over time.

Memories let you customise subagent system instructions, tailoring Phoebe’s reasoning to your team’s context, naming conventions, and preferred investigation style. With Memories, Phoebe goes beyond learning from past incidents and investigations — you decide what Phoebe should remember and how that shapes her future investigations and responses.

You can add memories manually to your list, or create them directly from investigations, capturing useful context as it happens.

🔁 Follow-Up Across UIs

Following last month’s launch of multi-turn investigations, Phoebe now keeps follow-ups in sync across interfaces — so whether you ask in Slack or from the web UI, you’ll see the same investigation threads and follow-up questions.

If you start a follow-up in Slack, it’ll automatically appear in the web UI; if you then continue an investigation with question in the web UI, the question and response will also be published back to the original Slack thread.

⚙️ New Subagent Architecture

We’ve started rolling out components of a new subagent architecture, designed for more effective collaboration between specialised agents.

It’s already showing strong performance gains in internal benchmarks — and will be fully available to users over the next couple of weeks. You’ll start noticing Phoebe investigating more complex problems with improved accuracy.

Customer love

❤️ Awesome updates. Really excited for these! Thanks for being so quick to respond and iterate

❤️ This is a huge UX improvement!

That’s the buzz for now — we’ll be back soon with more.
Phoebe team