Activate AI SRE separately; it does not require an On-call subscription or a member license. Every plan can create IM applications for AI conversations. Using an application for On-call notifications, including escalation policies and on-call reminders, still requires an active On-call Pro plan.
Overview
Incident response usually happens right inside your IM group — alerts come in, the team discusses in the channel, and someone spins up a war room. AI SRE puts the agent directly into that collaboration chain: you never have to switch to the console. You can summon it from IM to investigate, and your teammates can follow every step of its analysis in real time. AI SRE’s IM integration has four trigger modes:@Mention Summon
@mention AI SRE in a group chat or DM and describe the problem to start or continue an investigation. It replies in-thread, and the conversation context is tied to that IM session.
War Room Auto-Diagnosis
When someone opens an IM war room for an incident, AI SRE runs a preliminary diagnosis and posts the findings to the war room if the account has an enabled War room created automation rule that the incident’s channel and severity match. No one needs to summon it.
Automation Replies in the Group
Ask AI SRE in a group to create an ordinary automation rule and it uses this group as the run location: when the rule later fires on a schedule, an API call, or an On-call incident, the run happens in the original group and posts its conclusion there. Which session the run uses depends on whether that IM integration turns on AI SRE uses thread replies in regular group chats: in a direct chat and in a group where it is off or the platform doesn’t offer it, the run continues the agent session that chat already has, so the conversation so far stays in context, and a session is created only when that chat has none yet; in a group where it is on, each run opens a new session and a new thread, carries no earlier group conversation, and its first real output (progress or conclusion) is the thread’s root message, with no “started” placeholder posted before there is content; a run in a war room opens its own session too. The binding and scope are fixed at creation and cannot move to another group: before each run they are re-checked, and if the group’s team scope or session has changed the whole run is skipped.
Proactive replies
With Allow AI to reply proactively enabled on the integration, AI SRE also reads group messages that never mention it, and answers only when it judges a message worth answering; messages that @mention it always get a reply. Off by default, available only on Slack, Lark / Feishu and Microsoft Teams, and any single group can turn it off on its own.
Supported IM platforms
AI SRE’s IM interaction covers four major platforms. Each platform supports inbound @mentions (webhook), historical message reads, and outbound replies:
A “processing receipt” is the immediate signal the bot gives once it has your message; how it looks on each platform is described under Processing Receipt. Microsoft Teams also supports IM interaction and proactive replies (see Microsoft Teams), and is not in the table above; DingTalk and WeCom do not support proactive replies.
IM interaction requires that you have already connected the corresponding platform’s bot in Flashduty (the same IM bot used for alert notifications and collaboration). Complete the bot configuration in Flashduty’s IM integration first — only then can AI SRE send and receive messages on that platform.
AI SRE settings in IM integrations
Open the corresponding IM integration under On-call → Integration Center → Integration List → Instant Messaging and configure its AI SRE behavior under Advanced Features. Of the three switches below, the first two are on by default and the last one is off by default:- Allow AI SRE conversations in IM: available without enabling War Room. It controls both entry points — group @mentions and direct chats with the bot; in Slack it also controls the
/fdcommand. Turning it off makes group @mentions and direct chats with the bot (and, in Slack,/fd) silent. In WeCom’s Third-party App mode, only direct chats with the bot are available; group @mentions require Custom App Integration mode. - AI SRE uses thread replies in regular group chats: available only for Feishu/Lark and Slack. When enabled, each normal-group thread has an independent session; War Rooms still receive replies directly in the group.
- Allow AI to reply proactively: available only on Slack, Lark / Feishu and Microsoft Teams; DingTalk and WeCom have no such switch. When enabled, AI reads messages in the connected group chats that do not mention it and replies only when it judges a reply worthwhile; @mentions always get a reply. A message it reads but judges not worth answering gets neither a receipt nor a reply. Each group can opt out with
/ambient off(see /ambient command).
If AI SRE is not enabled for the account, these switches do not take effect. After AI SRE is enabled, each IM integration follows its own settings.
@Mention summon
In a connected IM group with Allow AI SRE conversations in IM on, @mention the bot and type your question, for example “@AI SRE check why the payment service’s 5xx rate is spiking.” The platform webhook forwards the message to AI SRE:1
Detect the mention
The platform distinguishes between group @mentions and direct messages. Once mentioned, AI SRE puts a “processing” receipt on that message (see Processing receipt) and checks whether the message contains a slash command.
2
Bind to a session
Multiple @mentions within the same IM chat continue the same agent session, preserving context. For a new session, a reply inside a thread brings in that thread’s starter message and a direct chat brings in that chat’s earlier messages, plus any associated incident as context. A top-level @ in a group brings only itself and the message it quotes (replies to), together with that message’s attachments: to include an error text or screenshot posted earlier in the group, reply to or quote that message. A reply inside a topic reads only that topic.
3
Read context and reply
AI SRE reads the thread / chat history to build context, investigates autonomously, and posts its findings in-thread. Other people @mentioned in the original message are carried over into the reply, making it easy for multiple parties to collaborate in the group.
Attachments and images
AI SRE downloads images, documents and voice messages from the chat and hands them to the agent as files. Attachments come from three sources: the message that @mentions it, earlier messages in the same thread or direct chat, and the message it quotes (replies to). A top-level @ in a group does not look back through the group’s earlier messages. To carry an error text or screenshot posted earlier, reply to or quote that message before the @mention. Inside a topic or a direct chat there is no such limit: post the error screenshot, then add “@bot what is this?”, and the model receives that screenshot. Per turn, the attachment quota is counted by type. The current message’s attachments take quota first, and history fills the rest, newest first:- History attachments that don’t fit are marked “not attached” in the prompt, so the model is never promised a file it doesn’t have.
- Rejected: an attachment over 20 MiB, or a direct link that is not HTTPS.
- Time budget for history: history attachments not downloaded within about 5 seconds are treated as not attached, so a turn never waits on an old screenshot.
- All attachments of the current message fail: AI SRE does not call the agent and replies with a short failure notice asking you to retry or send a smaller file. A failed history attachment does not stop the turn.
Processing receipt
Once the bot has your message, it puts an emoji reaction on the message you sent as a “received, working on it” signal (WeCom has no emoji API, so it posts a placeholder message that the task checklist card and the final answer then replace in place). The reaction is removed once the answer is delivered. Slack also shows a native “is working on it” status. With Allow AI to reply proactively on, a receipt is added only to the message AI decides to answer: a message it read but judged not worth answering gets neither a receipt nor a reply.How a reply reaches the chat
Once you hand AI SRE an investigation, every message that appears in the group is one of two kinds:- Result: an answer to your concrete question, or an actionable interim finding.
- Progress: a one-or-two-sentence update while work is still going on: confirmed progress, what it is waiting on, a real blocker. A progress message is not the answer: the result is normally posted as a separate message, and a turn can also end after a progress note when the model decides it owes you nothing more — that is not a failure.
flashduty safari send (see CLI): console sessions still answer with streaming text, and war-room auto-diagnosis is an unattended background turn that also keeps native streaming output.
When a reply is too long, AI SRE trims it or publishes the detailed evidence as an artifact and posts only the conclusion and the artifact link (see Artifacts). This limit does not apply when you explicitly ask for the full report in the chat.
War room auto-diagnosis
Whether AI SRE steps into a war room on its own is decided by the account’s War room created automation rules, not by a switch on the IM integration. A matching rule is what makes the diagnosis run. No one needs to @mention it, but you do need a rule in place:1
Create the war room
Create a war room (Lark / DingTalk / WeCom / Slack group) as part of the incident collaboration workflow.
2
Match a war-room-created rule
The platform matches the new war room against the account’s enabled War room created rules: the incident’s channel must be in the rule’s list and the incident’s severity must match the rule. Both have to hold before a run starts. With no matching rule, no analysis is produced.
3
Post the findings
The run is carried on that war room’s IM group (no console session is created). AI SRE investigates with the incident’s context and posts its findings back to the war room. Before anyone has started investigating manually, the first-cut analysis is already in the group.
War room auto-diagnosis is bound to the incident context. When the investigation begins, the platform attaches the
incident_id to the run, so AI SRE can read the incident details, timeline, and recent changes. For details on how incidents / war rooms interact with A2A, see Agent · Incident & War Room Integration.Standing tasks and the monitoring card
Sometimes the agent’s turn ends while a standing process-type task it started is still running, such as a monitor watcher or a background command. AI SRE then posts a monitoring card to the chat:While any standing task is alive, the IM session’s root message stays open, and so does the incident view. Later turns that task notifications trigger keep posting into the same chat. You don’t need to @mention the bot again.
Progress messages during a long investigation
In long investigations, AI SRE posts a progress update or two in the group saying what it is still checking, what is blocking, and what comes next. A progress message is not the answer, and the result is normally posted as a separate message; the turn can also end after a progress note rather than add a conclusion for its own sake. If nothing has changed, or it is waiting for your input, AI SRE may stay quiet.Connections and authorization
In the console, a missing credential shows an authorization card (see Console conversations). IM and API channels have no cards, so the agent handles it this way:
On the automation channel, nobody is watching and no reply will arrive. The agent doesn’t wait or ask questions. It records the missing connection or authorization as a blocker in its final report and delivers what the evidence already supports.
In-session switch commands
IM sessions have no picker UI like the console. Instead, slash commands let you adjust session bindings and proactive replies mid-conversation. Console sessions don’t support these commands: their environment and team bindings are fixed at creation, and proactive replies are configured only in IM groups. On Slack, write a command as/fd <command>, for example /fd stop.
The next three switch bindings and proactive replies.
/env — Switch Runner environment
Send/env <target> to the bot in a group or DM to rebind the current IM session to a different execution environment. The conversation history and context are preserved in full. After the switch, you can’t access the previous environment’s working directory. Files created or changed there during the conversation are gone. AI SRE re-stages skill and knowledge files in the new environment on demand.
<target> takes three forms:
Send
/env with no argument and the bot lists every target currently available: shared environments, your teams’ environments, and the cloud sandbox (the default sandbox plus any named templates). If a named sandbox template is not found, the bot replies “No cloud sandbox template named “x”. Use /env sandbox for the default sandbox.” along with the list of available templates.
The switch takes effect after the current turn completes and before the next message is processed — it does not interrupt a tool call that is already running. If the target runner is offline or not found, the command returns an error immediately and the binding remains unchanged.
On-premises deployments have no cloud sandbox. On-premises installs have no cloud sandbox infrastructure, so
/env shows no sandbox option or entries. You can only switch between BYOC runners and Auto./scope — Switch team scope
Send/scope <team-name> (or /scope personal to return to personal scope) to rebind the current IM session to a different team without ending the conversation. Team binding is a usage operation: any account member can bind a session to any team in the account. This fits war rooms, where responders often aren’t members of the incident’s team.
After the switch, AI SRE:
- Refreshes the memory snapshot for the new team scope (memories from the previous team scope no longer apply to this session).
- Queues the new team’s knowledge bundle to be mounted and injected into context when the next message is processed.
- Re-checks whether the session’s pinned execution environment is still available. If the pin is a BYOC runner or cloud sandbox template that belongs exclusively to another team, the bot resets the environment to Auto and replies ""x” does not belong to the current team. The runtime environment has been reset to Auto.” Shared and account-level environments are unaffected. If the check is inconclusive, for example because the template list can’t be fetched, the pin stays as is.
Knowledge bundles already mounted into this session are not removed — mounts are conversation-scoped, so switching back to a previously-mounted team will not duplicate the mount reminder.
When the group’s team or environment is unusable
A group conversation keeps the team and runtime environment chosen when it was created. An investigation can’t start when the group has no team or is set to a personal scope, the team was deleted, the pinned runner is offline or gone, the cloud sandbox template was deleted, the environment belongs to another team, or the deployment has no cloud sandbox. AI SRE then doesn’t start an investigation. It says whether the team, the runtime environment, or both are the problem, and lists what you can change: pick another team, pick another runtime environment, or go back to Auto. You can use/env or /scope, or say in the conversation which team or environment to use. Once the new setting is saved:
- If your original question had not started running, AI SRE runs it again automatically and posts the conclusion to the group, so you don’t need to resend it.
- If it had started running, it is not replayed, and the next message runs on the new configuration.
/ambient — Control whether this group gets proactive replies
Send/ambient <value> to the bot in a group chat to change how that one group gets replies, without touching the integration switch or any other group:
Sending a bare
/ambient is the same as /ambient status.
- Allow AI to reply proactively on the integration is the master switch, and a group’s value only works within what it allows: with the integration switch off,
/ambient onhas no effect, while/ambient offalways works. - It can only be configured in a group chat; sending it in a DM is refused. DingTalk and WeCom do not support the command.
- A group’s value is kept for 30 days, after which the group goes back to following the integration switch.
- Whatever the group’s value, a message that @mentions the bot always gets a reply.
- On Slack, the same action is
/fd ambient <value>.
/feedback — Rate an AI SRE reply
Once an AI SRE reply lands in a group chat, anyone in the group can rate it — not only whoever started the investigation. Send/feedback to the bot to rate or comment on the reply you are looking at:
- Entry and resolution:
/feedbackapplies to the reply you are looking at, even if you never took part in that conversation. - Eligibility: not limited to session members — any account member who can use AI SRE may rate (read-only group-chat readers included).
- How to rate: choose
like/dislike/none. A comment-only submission (no rating) is allowed and does not wipe an earlier rating you gave, while an explicitnoneclears the existing rating. - Several raters: when several people rate the same reply, each rating is kept and none overwrites another.
- Console side: the console conversation header also offers thumbs-up feedback on replies (see Feedback), sharing the same rating system as IM
/feedback.
Relationship to console sessions
Whether you start a session from IM or the console, it’s the same AI SRE. It handles messages one at a time and streams output. It compacts context automatically in long conversations and supports optional team and environment bindings. The only difference is the entry point —- Console: ask questions one at a time in the Chat workspace and view the full tool-call and sub-session panels. Environment and team bindings are fixed at creation and cannot be changed.
- IM: @mention it in the group where your team already works; findings are posted back to the thread. Use
/envand/scopeto switch bindings mid-conversation,/ambientto control whether the group gets proactive replies, and/stopto stop the current analysis at any time. Great for grabbing a quick analysis at the incident scene, then diving deeper in the console.
Related pages
Overview
Learn about AI SRE’s overall capabilities, typical use cases, and console navigation.
Console Chat
Learn about sessions, streaming output, cancellation, and context compaction.
Agent
Inbound Agent Card and incident / war room integration.
Usage Insights
Use /insight to review the past 30 days of sessions and surface operational friction.