As federal agencies look to implement agentic artificial intelligence to speed up workflows in 2026, AI infrastructure providers say data organization is a critical first step to making agentic solutions effective.
Public sector leads at AWS, Oracle and Cisco said clients want to pivot away from chatbot capabilities to focus on more actionable investments in AI systems, including domain-specific models that can handle specific tasks, according to a Nextgov/FCW report. To get there, however, data must be structured to be made actionable.
“Agentic AI is where the industry is headed, is where our customers are headed. It’s where they’re demanding outcomes,” Rishi Bhaskar, director of public sector partner sales at Amazon Web Services, told Nextgov. “But that starts in the data journey.”
Kapil Bakshi, a distinguished engineer within the Office of the Chief Technology Officer at Cisco’s U.S. Public Sector branch, said that “the sentiment among government technology leaders has shifted from ‘what is possible’ to ‘what can we operationalize.’”
Bhaskar and Bakshi concurred that delivering agentic AI solutions requires both modernizing legacy infrastructure and code, as well as refining and organizing data to make it easy to turn into actionable information.
“We have tremendous opportunity for our civil servants, our partners … to actually get away from manual, repetitive tasks and shift to the value at creation that we all like to do,” Bhaskar told the publication.
At Oracle, 2026 will bring a focus on updating data assets to enable “context-aware AI,” according to Peter Guerra, vice president of Data and AI for Government Defense and Intelligence. “The approach that Oracle has always taken, because we are a data-first company, is: AI that knows your data is the only useful AI out there,” he said. “What we are looking forward to from our customers working specifically in the federal space is a continued path of modernization of that data asset to enable context-aware AI.”
The specific workflows the three companies will look to automate with agentic AI in 2026 vary, according to Nextgov, and include network traffic management, data entry and document review.
Cloud-computing infrastructure is another key deliverable for scaling data-dependent agentic AI applications, the executives said.
“If you look from a commercial standpoint where Oracle Cloud is being used, we’re used very heavily in the AI startup space and the frontier model companies and so forth,” Guerra said. The cloud “gives customers the flexibility they need, from a consumption standpoint, to consume just what they need.”
“Speed and acceleration, I think, [are] also important,” Bhaskar added. “The faster we can get the cloud, the faster we can unlock innovation.”
