Artificial intelligence is hitting an inflection point as agentic AI pushes beyond copilots and into the core machinery of enterprise operations.

What used to be simple assistance is evolving into a distributed network of agents that need consistent data access, deterministic behavior and deeper security baked into every layer. The pressure is on: Companies must harmonize identity, observability and lifecycle governance while preparing for systems that think, respond and coordinate in real time. It’s a dramatic shift that’s reshaping expectations for how modern platforms should actually work, according to Paul Nashawaty (pictured, right), principal analyst at theCUBE Research.

“When we look at entering the agentic AI era, we see that agents are evolving from assistance or copilots to really first-class participants in the enterprise workflows,” Nashawaty said. “We’re also seeing that complete AI lifecycles are really getting framed up, and we saw this at Ignite at the keynote.”

Nashawaty spoke with fellow analyst Rob Strechay (left) as part of an AnalystANGLE during Microsoft Ignite, for an exclusive conversation on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how Microsoft Corp. is advancing agentic AI through unified data, lifecycle tooling, security foundations and next-generation productivity agents. (* Disclosure below.)

How agentic AI is shaping the next phase of cloud and data platforms

Organizations are rapidly adopting tools that unify data, identity and operational telemetry as agents take on more autonomous responsibilities. This shift is driving demand for platforms that handle everything from lifecycle governance to real-time observability while reducing fragmentation across productivity, data and security layers, Nashawaty noted.

He pointed to the Ignite keynote as a clear example of that momentum, saying, “It reinforced Microsoft’s AI lifecycle management and messaging along from AI pilots into production. We’re seeing in our own CUBE Research … that 51% of applications that are going into production are using AI today.”

At the same time, data infrastructure is becoming a competitive battleground for agentic AI, especially as companies aim to eliminate redundant pipelines and avoid recreating data across multiple environments. Enterprise teams are prioritizing open formats, zero-copy interoperability and consistent metadata to maintain trust and reduce operational drag. These trends are fundamentally reshaping how organizations build intelligence layers atop increasingly complex data estates, according to Strechay.

“Microsoft was introducing zero-copy data interoperability, enabling organizations to really have data across partner data platforms and Fabric OneLake with no data movement, which is really huge to them,” he said. “Again, they are looking to help companies with analytics, BI and AI workloads that can run on either platform.”

Security’s new role in the agentic AI stack

Security is emerging as one of the most consequential layers in the agentic AI landscape, with autonomous defenses and identity-aware controls running across cloud, data and collaboration systems. Enterprises are embedding guardrails directly into workflows to prevent drift, oversharing and permission sprawl as both human and machine identities grow more intertwined. This reinforces the critical need for security architectures that evolve alongside agentic behavior, according to Nashawaty.

“Security as a primitive … ambient and autonomous security is built into every layer,” he explained. “You start at silicon, which is the hardware and move to the operating system, move to identity management, that could be human identity, machine identity, non-human identity, cloud, data, apps and agents. All of this is a big part of the security angle.”

Agentic AI is also redefining productivity experiences, extending beyond traditional chat interfaces into workflow execution, multi-step orchestration and domain-specific agents. Companies are increasingly adopting agents that can act on a user’s behalf in context-rich environments, reducing friction and expanding access to AI-powered workflows across the organization, Strechay emphasized.

“These agents weren’t just about asking questions and being chatbots, they actually can act on your behalf as well,” he said. “They’re also a new pattern where customers could really use chat to drive multi-step AI-powered workflows and they’re calling this agent mode.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of Microsoft Ignite:

(* Disclosure: TheCUBE is a paid media partner for Microsoft Ignite. Sponsors of theCUBE’s event coverage do not have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE