Skip to main content
Resolve for incident channels captures context directly from your incident channels, and brings autonomous investigations into your incident workflows. It builds an evolving picture of the incident, and helps responders move faster at every stage - from first detection to postmortem.

Feature Overview

Incident channels let ResolveAI treat specific chat channels as live incident investigations:
  • Recognizes incident channels by naming convention Admins define one or more channel name patterns (e.g. incident-*, *-sev?, inc-*-sev-*). Any channel Resolve is invited to whose name matches a pattern is treated as an incident channel. Available in Slack, and in Microsoft Teams when you use Channel per incident recognition.
  • Captures and understands incident context Resolve understands channel messages, links, logs, and other artifacts. It classifies them into symptoms, human findings, impact updates, and evidence so it can reason about the incident as it unfolds.
  • Runs an autonomous investigation in the background Once it has enough context to understand the problem, Resolve starts an incident investigation and continuously updates it as new information appears in the channel. As the investigation progresses, Resolve can also proactively share evidence-backed findings directly in the channel to help responders get to the root cause faster (Slack only).
  • Answers questions with full incident context Any @-mention in the incident channel sees the current investigation state, including prior messages, links, and evidence. Responders can ask Resolve to check logs, dashboards, inspect deployments or PRs/MRs.
  • (Slack only) Provides instant “catch-up” summaries for new joiners When a new person joins the channel, Resolve can send them a private, concise summary of what’s happened so far—only visible to them.
  • Generates postmortems Resolve can create a postmortem draft from the incident channel’s activity and the investigation: timeline, findings, impact, and supporting links, ready to paste into Google Docs, Notion, Jira, Linear, and more.

Getting Started

Prerequisites:
  1. Resolve’s App for Slack is installed in your workspace.
  2. You have administrative permissions to workflows in your incident automation tool.
  3. You have Admin access to ResolveAI.

1. Set up how Resolve recognizes incident channels

In the left panel, go to Admin > Incident Channels. Under Set up, you’ll configure two independent recognition paths. Use either or both:Auto-join webhook (recommended)For public incident channels your paging tool opens. When your tool posts a webhook to Resolve, Resolve joins the channel and treats it as an incident automatically (no name pattern required).
  • Select the collaboration tool integration instance (e.g.: mySlack)
  • Create a new webhook
  • Copy the webhook endpoint URL and headers to your incident orchestration tool (e.g.: PagerDuty, Opsgenie, Zapier or a custom incident Slackbot)
  • Configure the webhook payload as per the instructions (e.g.: {"channelId": "<channelId>"}...)
This automation adds Resolve to public channels only, because Slack doesn’t allow apps to add themselves to private channels. For a private channel, use the pattern path below and invite Resolve manually.
Channel name patternFor channels a human adds Resolve to (/add the Resolve app, or @Resolve in the channel), public or private. Any channel whose name matches a pattern is treated as an incident.In the Channel name pattern card, add one or more patterns. Use * to match any run of characters and ? to match a single character. Examples: incident-*, *-sev?, inc-*-sev-*.

2. Decide how Resolve acts

In the Decide how Resolve acts section, choose how Resolve behaves. Settings apply org-wide to every incident channel Resolve joins:
  • Respond when mentioned
    • Resolve responds when explicitly @mentioned in the channel.
    • This is always on, ensuring Resolve is available in the channel
  • Catch up new participants (Slack only)
    • Toggle this on so new joiners get a private summary of the incident so far.
    • Recommended for busy incident channels with many stakeholders.
  • Proactively share insights (Slack only)
    • Toggle this on to let Resolve participate proactively in the channel in two ways:
      • Thread follow-ups: In a thread Resolve is already participating in, it understands follow-ups without a new @mention (e.g. “can you elaborate on that?”) and responds. It won’t chime in on threads it isn’t already part of.
      • Incident insights: When Resolve’s investigation finds something worth flagging, it shares a concise summary in the channel. These are evidence-backed, specific, and haven’t already been discussed. The goal is to surface new, concrete signals that help responders get to the root cause faster, not repeat what’s already known. Posting mode for incident insights controls how Resolve posts:
        • Resolve decides (Recommended): Resolve posts new messages for new working theories and root cause shifts, and replies in-thread for supporting details and follow-ups.
        • Single thread: Resolve posts all updates only within the thread it creates at the start of the investigation. Anyone following that thread will receive notifications as new updates are added.
Note: Proactive messages include a Mute button. If someone mutes it, Resolve will stop posting proactively in that incident channel.

3. Use Resolve in your next incident channel

Once the above is configured:
  1. An incident is declared via your chosen workflow (e.g. PagerDuty, Slack command, Zapier, Opsgenie, etc.).
  2. A channel is created with a name matching one of your incident channel patterns, and Resolve is invited using a webhook.
  3. Resolve captures incident context, starts an incident investigation, and posts an “Investigation in progress — I’m here to assist.” message.
  4. Responders:
    • ask @Resolve to check logs, deployments, impact, or status
    • use catch-up summaries for new joiners
    • request a customized postmortem draft at the end of the incident by using an inline prompt such as: @Resolve create a post mortem for me that covers: 1. Incident title, 2. A summary and narrative of the issue, its impact and the root cause, 3. Details of the root cause, 4. Detailed timeline of findings and actions, 5. Impact analysis, 6. Lessons learned and action items