Cloud observability platform today announced that it is launching a generative AI model to augment its existing machine learning tools, which focus on a company’s observability data, with additional information from other data sources. Through this, the tool can now generate smarter remediation recommendations and, in turn, improve ops teams’ reaction times.

The company, which focuses on aggregating data from a variety of open source observability tools, started using supervised machine learning back in 2016 and then expanded that with Cognitive Insights, which draws from crowdsourced information from the community and the other forums, social threads and open source repositories.

“Up until now, Cognitive Insights generated crowdsourced recommendations,” said CTO Asaf Yigal. “There’s a high likelihood that, among those billions of events, there’s useful and important information, and the system attempts to determine which links or discussion forums might provide additional context to help remediate the issue at hand. But with the use of ChatGPT, we’ve added the ability to make more precise recommendations based on crowdsourcing, offering potential solutions and additional investigative paths to the problem. In short, we can now deliver a more accurate recommendation.”

Yigal stressed that has a pretty unique dataset, which includes its users’ search behavior and metadata. Combined with OpenAI’s models, the service can now generate detailed — and contextual — remediation advice.

“As opposed to other ML technologies, the main difference is access to contextual data about how engineers investigate issues across millions of software libraries and products,” he said. “This access to proprietary data is what makes this new system so fundamentally different. In addition, as more people use it, the system becomes smarter. Generic responses are useless while specific contextual advice can be a game changer.”