As organizations race to operationalize artificial intelligence, application programming interfaces are becoming a hotbed of discussion. It is increasingly clear that the right API platform decides which enterprises lead and which fall behind.

Agentic patterns demand real-time context, safe action-taking and tight feedback loops; all three break down if the substrate of APIs, identity and catalogs isn’t in place. This once-speculative convergence of AI and APIs has now arrived in full force , according to Reza Shafii (pictured), vice president of product at Kong Inc.

Kong’s Reza Shafii talks with theCUBE about API platform management.

“I think the convergence of AI — which obviously that’s the big talk — and APIs is really coming to reality,” he said. “I think people are feeling that connection between the two is really important.”

Shafii spoke with theCUBE Research’s Paul Nashawaty at the Kong API Summit: The API Summit for the Agentic Era, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed Kong’s newly-announced product lineup how it fits into enterprise API management strategies. (* Disclosure below.)

Agentic AI thrives on a solid API platform foundation

Before teams debate tooling, the first-order problem is capability: large language models are smart, but they can’t act or fetch context on their own. That makes resilient, discoverable APIs the precondition for reliable agent behavior, according to Shafii.

“An LLM is brilliant on its own, but it can’t do things. It’s very hard for it to retrieve on-the-fly, real-time information,” he said. “Long story short, if the LLMs are the brain, [then] the APIs are the limbs. … If they’re not solid, guess what? The AI is going to misbehave. The brain is going to misbehave.”

Once that foundation exists, the next hurdle is scale: reducing complexity and closing skill gaps so more teams can build safely. Kong positions Model Context Protocol and Kong Konnect as tools to harmonize API and LLM calls and simplify composition, Shafii explained.

“You can actually weave together backend APIs — multiple tools into a single tool,” he said. “Let’s say you have an API for all of your product usage data, an API for your actual customer support information and then an API for your product requests. You can create a tool on the fly that says all of these APIs should be a tool that tells you the customer sentiment. That would take I would say 20 to 30 minutes … to create and make available on Konnect.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the Kong API Summit event: 

(* Disclosure: TheCUBE is a paid media partner for the Kong API Summit. Neither Kong, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE