There have been a range of efforts in recent years to take advantage of artificial intelligence for information technology operations, or AIOps. Using machine learning, these initiatives have enjoyed limited success at correlating alerts and detecting anomalies across the vast IT telemetry streams. Now, AI for IT operations is poised to deliver on its long-promised potential — but only if enterprises learn from past mistakes.

Today, many organizations are taking advantage of agentic AI by adopting isolated agents or copilots. However, this approach creates a fragmented structure that recreates the very siloes AIOps was meant to eliminate, leaving enterprises hesitant to roll out these solutions into production, according to Bhaskar Krishnamsetty (pictured, left), chief product officer of Fabrix.ai Inc.

Fabrix.ai’s Bhaskar Krishnamsetty, distinguished engineer Rached Blili and chief AI and marketing officer Shailesh Manjrekar speak with theCUBE about building and scaling their agentic AI platform.

“There is a lot of excitement from our customers to embrace agentic AI, but at the same time, there is some hesitation — and also fragmentation [in] how they’re approaching [it],” Krishnamsetty noted. “Either they try isolated agents, or try some [large language model] copilots. The hesitation comes when they want to operationalize the production environment to scale.”

Krishnamsetty, Rached Blili (center), distinguished engineer at Fabrix.ai, and Shailesh Manjrekar (right), chief AI and marketing officer of Fabrix.ai, spoke with theCUBE’s Bob Laliberte  and Zeus Kerravala, principal analyst of ZK Research, a division of Kerravala Consulting, at the Agentic AI Unleashed: The Future of  Digital & IT Operations event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed the gaps found in both traditional approaches to AI for IT operations management and how agentic AI could bridge those concerns. (* Disclosure below.) 

Behind AI for IT operations and the ‘holy grail’ of AIOps

Early AIOps solutions required significant manual effort, requiring months to connect to all data sources, Krishnamsetty noted. While functions such as data integration, correlation and anomaly detection have matured and become more autonomous, downstream use cases still require a hands-on approach, he explained.

“Whether it is … root cause analysis or auto remediation, it’s manual and semi-automated,” he said. “The true promise, the holy grail of AIOps — from ingestion to detection to resolution — [is that] now with these AI agents filling those gaps, our customers can really take that next step.”

In Fabrix.ai’s answer to AI for IT operations management, individual agents — whether for security, cloud or financial operations —  work as intelligent digital operators, continuously gathering telemetry, interpreting system state and executing corrective actions while keeping humans in the loop. The SecOps agent illustrates the power of this approach, according to Krishnamsetty.

“It goes and reads through what the exposure is and … what configuration it is vulnerable,” he said. “It scans the entire environment and logs in to those particular devices and checks the configuration. If it finds any vulnerability, the last part is customizable. Either it can create change requests or it can automatically take corrective actions. We can put the human in the loop as well.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the “Agentic AI Unleashed: The Future of Digital & IT Operations” event:

(* Disclosure: TheCUBE is a paid media partner for the “Agentic AI Unleashed: The Future of Digital & IT Operations” event. Neither Fabrix.ai, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE