A human-aware change agent that has been created to use the insight of people to drive automated incident investigation has been released by software delivery platform provider Harness as a part of its AI SRE unified response system.
While most AI approaches to software reliability involve the use of logs, metrics, traces and anomaly detectors, Harness said the reality of incident response is that most of the important clues about failure come from people reporting problems. Those human reports involve such clues as “the customer said the checkout box froze right after they updated their cart,” or “service X felt slow an hour before this started.”
According to the company, the Human-Aware Change Agent can listen to the conversations engineers have over Slack, Teams or Zoom, for example, the uses those insights into actionable intelligence that drives automated change investigation. The way it works is that it uses Harness’ AI Scribe tool to record those conversations and the agent uses the human signals to direct the automated investigation across changes that include deployments, feature flags or changes to configuration, infrastructure updates and ITSM change records.
