Enterprises are struggling to move artificial intelligence beyond the testing phase. The next step in this evolution requires new, purpose-built agentic AI platforms designed for governance, trust and security — not just simple automation.

Building these platforms means looking beyond the technology that’s captured most of the headlines. Though powerful on their own, large language models are only one piece of a larger puzzle, according to Shailesh Manjrekar (pictured), chief AI and marketing officer of Fabrix.ai Inc.

Fabrix.ai’s Shailesh Manjrekar speaks with theCUBE about building and scaling an agentic AI platform to move enterprise AI from isolated demos to production operations.

“To us, LLMs are just a means to an end,” Manjrekar said. “You’ve got to have the right building blocks around it, and you really have to have a platform which caters to … governance, observability, explainability, trust and security.”

Manjrekar spoke with Bob Laliberte, principal analyst at theCUBE Research, and Zeus Kerravala, principal analyst at ZK Research, a division of Kerravala Consulting, at the “Agentic AI Unleashed: The Future of Digital & IT Operations” event, during an exclusive interview on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed using agentic AI platforms to move between isolated AI demos and scalable, autonomous enterprise operations. (* Disclosure below.)

Solving the ‘last mile’ with agentic AI platforms

While enterprises might aspire to turn insights into action, they may not be equipped to build the necessary AI, data and automation pipeline. The challenge lies in scaling from a simple proof of concept to a complex production environment, Manjrekar explained.

“It’s one thing to demonstrate a demo with one use case [and] with a few operational assets,” he said. “It’s another thing to be getting data from 500 different assets and being able to cater to bespoke use cases. That’s where the platform approach really stands out.”

This approach is designed to solve the persistent “last mile issues” seen in previous implementations of AI for IT operations, or AIOps. By creating productivity gains, building out use cases then consolidating practices, data silos can be broken down, according to Manjrekar.

“You have these agents where are you are using LLMs as a judge, you’re using LLMs as agents, to actually create [these] multi-agentic orchestration workflows, which can take care of a lot of operational tasks,” he said. “That’s where the speed and confidence comes in.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of at 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