It support · Production

incident.io builds cooperative AI agents for incident response to avoid ironies of automation

The problem

As software is increasingly built and operated by AI, incident responders have diminishing context about the systems they manage. Automation designed to help can paradoxically atrophy human skills, leaving responders underprepared when automation itself fails.

First attempt

Over-automated approaches risk turning human responders into passive rubber-stampers who confirm AI decisions without genuine understanding, and AI acting without full operational context can escalate minor incidents into full outages.

Workflow diagram · grounded in source
1
Incident fires and pages responder
trigger
“An incident kicks off. Services are falling over, people are scrambling”
2
AI retrieves logs and change context
ai_action
“The logs show an uptick in 'timeout' errors starting at 14:23, which coincides with this config change. If that looks correct, you can page the team by clicking below”
3
Pattern match against past incidents
ai_action
“This alert fired during incidents in Feb and Oct last year—both related to queue backlogs in service-X after deploys. Might be worth checking if that's happening again. See INC-2412 and INC-2508 for more details”
4
Agent surfaces hypothesis with reasoning
output
“it will show its working: what it looked at, why it thinks it's relevant, and where there's uncertainty”
5
Responder reviews and decides
human_review
“keeping humans in the loop—and responsible—when it comes to meaningful decisions”
Reported outcome

incident.io is building cooperative AI agents anchored to human-AI collaboration: surfacing context and hypotheses transparently, keeping humans responsible for key decisions, and amplifying rather than replacing human judgment, with tools like Scribe and Investigations already shipping.

Reported metrics
Incident resolution speedfaster incident resolution
Customer impactreduced customer impact
Responder cognitive burdenless cognitive burden
Reported stack
ScribeInvestigations
Source
https://incident.io/building-with-ai/avoiding-the-ironies-of-automation
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

incident.io is building cooperative AI agents anchored to human-AI collaboration: surfacing context and hypotheses transparently, keeping humans responsible for key decisions, and amplifying rather than replacing huma…

What tools did this team use?

Scribe, Investigations.

What results were reported?

Incident resolution speed: faster incident resolution; Customer impact: reduced customer impact; Responder cognitive burden: less cognitive burden (source-reported, not independently verified).

What failed first in this deployment?

Over-automated approaches risk turning human responders into passive rubber-stampers who confirm AI decisions without genuine understanding, and AI acting without full operational context can escalate minor incidents…

How is this it support AI workflow structured?

Incident fires and pages responder → AI retrieves logs and change context → Pattern match against past incidents → Agent surfaces hypothesis with reasoning → Responder reviews and decides.