Harness Inc., a software delivery startup that provides artificial intelligence tools for developers to update and monitor applications, said today it’s launching a “human-aware change agent” to help teams resolve incidents faster without adding extra members.

The company said AI site reliability brings together the critical components of response engineering by capturing context, coordinating action and providing investigative assistance. AI Scribe, the new product, was built to glue all that together.

When AI is used in SRE, a lot of journalists use words such as “surface” and “insights” because we treat analysts and engineers like detectives who are looking at incident reports, logs and diagnostics a lot like detectives overlooking a crime scene. This is a chaotic mess of information, often a brutal cascade of patterns that can take a lot to sift through without basic knowledge of not just the system but a foundational understanding of the what, where and why, and who reacted when.

Furthermore, sometimes the most important “clues” come from human complaints or responses early on. For example, they might be “The customer said the checkout button froze right after they updated the cart,” or “Service X felt slow an hour before this started,” or “Didn’t we flip a flag for the recommender earlier today?” or “This only happens in the US-East cluster.”

These kinds of messages that appear in Slack, email or other messaging systems can quickly shape the investigation long before anyone begins to glance at reports, logs or warms up a dashboard.

Harness said AI Scribe operates as a conversational interface that captures operational signals from the team’s natural dialogue surrounding impacted services, dependencies, customer-reported symptoms, reported theories or contradictions, key sequence-of-events clues (“right before…” and so on) and attempts to synthesize them.

It investigates these “clues” as human signals and provides a change graph using feature flags, configuration files, infrastructure updates and other elements natural to the structure of operational machinery. The aim is to provide a hypothesis that can be understood by the team.

Instead of throwing wild guesses at the team, it produces clear, explainable insights with reasoning and supporting data. This allows teams to quickly validate or discard ideas and move on.

Harness said it wants AI Scribe to act as an opportunity to become an incident responder that provides functional intelligence, providing a faster path to truth, not just summarization.

Photo: Pixabay