Where to Interact with Rootly AI
In an incident channel (most common)
Rootly AI reads details about your incident, makes updates, answers questions, generates comms, and pages directly in your incident channel (subject to your user permissions in Rootly).
Rootly AI is added automatically to incident channels created through Rootly. Get started by declaring an incident using the shortcut /rootly new.
Start chatting with Rootly AI by mentioning @Rootly in the channel or thread and asking “catch me up” or “page the database on-call team”.
Rootly AI always replies back in-thread.
Slack assistant side-pane
Best for: questions about a specific incident in a private conversation.
When there’s an active incident in your view, you can ask Rootly AI any question about the incident. Rootly AI can only answer questions here — it cannot take action on your behalf. This surface is best used when you want to catch up but don’t want to distract the responding team in the incident’s Slack channel.
Ask Rootly AI questions in the incident channel when the context might be relevant to other teammates, so they can also benefit from the reply.
DM Rootly AI directly
Best for:
- Quick on-call lookups: “Who’s on call for payments right now?”
- Looking up an incident you’re not actively in: “Catch me up on INC-1234.”
- Personal questions: “What action items are assigned to me?”
Rootly AI can only answer questions here — it cannot take action on your behalf. This surface is best used when you want to catch up but don’t want to distract the responding team in the incident’s Slack channel.
Access is still governed by your Rootly permissions, so you can only see incidents and data you’re entitled to.
Interacting with Rootly AI
Once you’ve asked Rootly AI a question (starting with @Rootly), you’ll see a “Thinking…” indicator as Rootly AI gathers context and builds a plan. These indicator messages rotate every few seconds as Rootly AI continues to work.
The thinking-indicator messages hint at what Rootly AI is planning to do next. For example, “Checking who’s on call” means Rootly AI is gathering on-call context for your question.
Giving Feedback
Use Helpful (👍) and Not helpful (👎) to rate Rootly AI’s replies. One click captures whether the answer was useful and helps Rootly continuously improve. Your feedback flows into Rootly’s evaluation pipeline and does not directly fine-tune the underlying model.
If you spot something Rootly AI got wrong, the most useful thing you can do is click Not helpful and provide additional feedback. That signal goes straight into our regression set.
Providing Clarifications
Clarification cards are the only structured place where Rootly AI pauses for your input on how to proceed. If something is ambiguous or required data is missing, Rootly AI will ask rather than guess.
Examples of clarifications include:
- Ambiguous person. “Page David.” → “Which David did you want to page?” with options to select the right one.
- Ambiguous severity. “Escalate this.” → “What severity should I set?” with buttons for the team’s configured severities.
- Missing required field. Moving to a status that requires a cause, environment, or other field — e.g. “What environment is affected?” — with fields to provide the necessary data.
Using Clarification Cards
Based on how much input Rootly AI needs, you’ll be able to reply via:
- Up to 5 options → rendered as buttons. The first button is highlighted as the primary choice.
- 5 to 25 options → rendered as a dropdown (single-select).
- More than 25 options, or a question without a fixed answer set → free-text mode; reply in the thread.
After You Respond
A small task pill appears in the thread reading User answered with the question and your chosen label — e.g. Which David did you want to page? → David Chen. Rootly AI picks up from where it left off.
Expiry
Clarification cards expire after 60 minutes. If you click an expired card, you’ll see:
⌛ This question has expired.
To continue the conversation after 60 minutes, simply re-ask your original question to get a fresh card.
Taking Action with Rootly AI
Actions like severity changes, status transitions, role assignments, and paging are executed immediately by Rootly AI when you ask for them. Rootly AI confirms what it did in its reply.
For example:
- “I’ll take commander.” → “You’re now the commander.”
- “Escalate to SEV1.” → “Set severity to SEV1.”
- “Page the SRE on-call.” → “Paged the SRE team via their escalation policy.”
If Rootly AI needs more information before it can run a write (for example, which David to page, which severity to use), you’ll see a clarification card first. Otherwise it proceeds, and if anything fails (permission denied, validation error, transient tool error), Rootly AI tells you in plain English in the same thread.
Every action is taken by your Rootly user, with your permissions, and is attributed to you in the audit trail. Rootly AI cannot perform any action you couldn’t perform yourself in the Rootly web app.
Examples
”Catch me up”
The most common ask. What you’ll see:
- Type
@Rootly catch me up in the incident channel
- The thinking indicator appears, then “Pulling up incident details…” → “Checking incident status…” → “Gathering incident data…”
- A short structured summary appears: what happened, current status, what’s been tried, who’s involved, what’s open.
- The Helpful / Not helpful feedback buttons sit at the bottom.
- Ask follow-ups in the same thread. Rootly AI remembers what it just pulled — “What caused it?” / “Who’s been paged?” / “Show me similar past incidents.” all work as continuations.
”Page on-call team”
This shows the clarification flow in action:
- Type
@Rootly page the on-call team in the incident channel. Status shows “Looking up schedules…” → “Checking who’s on call…”
- Rootly AI searches your Rootly catalog for a team matching “database”.
- If multiple teams match, Rootly AI posts a clarification card:
"Which team did you mean?" with a dropdown. You pick one and click Confirm.
- Rootly AI pages via Rootly On-Call to the team’s current on-call, using their configured notification preferences. Rootly AI replies:
"Paged the Rootly team via the Default Escalation escalation policy." The audit trail attributes the page to you, not to Rootly AI.
- If you don’t have permission to page in Rootly, Rootly AI tells you so and stops. The action runs as your user, so it can’t do anything you couldn’t do yourself.
