Video · 2026-04-24 · 20m · 6 moments

How to Adopt AI Agents in Your Enterprise with Google Cloud's Riyaz Habibbhai

✦ AI generated

timeline · colored by role

Claim

An integrated, full-stack AI platform (chips to models to applications) eliminates the integration challenges that are enterprises' number one issue, cutting implementation and adoption costs, while chip-optimized model training (e.g., on TPUs) lowers token costs that get passed on to customers.

Riyaz argues Google's full-stack integration (hardware through apps) removes the top pain point for enterprises—tool integration—while also lowering the cost of AI intelligence via chip-optimized model training.

transcript

Riyaz Habibbhai: What's the number one issue enterprises face is integration challenges, right? Having an integrated stack means out of the box it all works together. So, you're saving on implementation cost, you're saving on adoption costs, right? All of that is just savings for an organization.

explains mechanism · 1

Claim◆

Gemini Enterprise Agent Platform is not a rebrand of Vertex AI but a substantive evolution that layers major new governance, identity, security, and scaling capabilities (like cryptographic agent identity) on top of what Vertex already provided.

Riyaz explains that Vertex AI, long the de facto developer platform for agents, has been significantly enhanced with governance, security, and identity capabilities and renamed Gemini Enterprise Agent Platform to reflect that expanded scope.

transcript

Riyaz Habibbhai: Vertex AI for the last few years has been our de facto developer platform for you to build uh agents and applications and enroll it out within your organization. Now, over the last couple quarters, we've enhanced that platform significantly. We brought new capabilities into governance, security, identity for you to be able to run agents, right? Give agents a cryptographic ID so they are identified within the organization.

extends · 1

Claim

Governance, security, and compliance capabilities are included as standard, built-in features of the Gemini Enterprise app rather than a separately priced add-on, because organizations shouldn't have to pay extra to adopt agents safely.

Riyaz states that governance, security, and compliance are foundational and come standard with Gemini Enterprise, powered by the same runtime used across developer and end-user surfaces.

transcript

Riyaz Habibbhai: Governance, security, compliance, these are foundation for you to go agentic. And we fundamentally believe you shouldn't have to pay extra to go agentic, right? So, now when you adopt the Gemini Enterprise app to give the power of AI to every employee, those foundational capabilities just come out of the box, they're standard.

Mechanism

Openness means giving customers MCP endpoints for their data stores as a fallback to pre-built connectors, so they aren't blocked waiting on a vendor to build a specific connector before they can integrate their data into Gemini Enterprise.

Riyaz describes Google's 'MCP everywhere' approach, letting customers connect any MCP-enabled data store directly to Gemini Enterprise rather than waiting for a purpose-built connector.

transcript

Riyaz Habibbhai: Connectors are great and they work, but there's a limit to how many connectors you build because folks have very different data stores... having the ability for you to have an MCP endpoint for all your data store, you can then just work against that MCP endpoint. You don't have to wait for the connector, right?

explains mechanism · 1

Anecdote◆

Macy's built a fully voice-activated, multi-agent commerce and service experience on Gemini Enterprise's customer experience platform in just four weeks, going from nothing to production.

Riyaz cites Macy's as an early-adopter example, describing an orchestrated set of agents (search, transaction, service) built into a voice-activated front end in four weeks.

transcript

Riyaz Habibbhai: One of our customers that has gone fully into agent-driven commerce and service is Macy's... Macy's live built on the Gemini Enterprise customer experience platform went from nothing to a fully voice-activated agent on their website in 4 weeks, right? And that is not just a single agent. There's an orchestration of agents from searching the catalog to helping transact in commerce.

Claim

The main obstacle for enterprises adopting agents isn't the technology itself but choosing the right platform amid widespread 'agent washing,' where every vendor claims the same capabilities regardless of actual roadmap or delivered efficiency.

Riyaz argues that customers' biggest challenge is cutting through 'agent washing' to evaluate which platforms genuinely invest in and deliver agent capabilities at the right cost and efficiency.

transcript

Riyaz Habibbhai: There are a lot of conversations, there's a lot of quote-unquote agent washing that's happening out there, right? Everybody seems message seems to be the same. It's like we have the platform, we can do it. And the guidance I've given to customers is evaluate if it's truly a platform that is investing in the right direction, has the road map to continue to build in that direction, and is actually delivering you something at the cost and the efficiencies that needs to be delivered.

Highlight slides

Gemini Enterprise Agent Platform: Evolution, Not Rebrand✦ from: Gemini Enterprise Agent Platform is not a rebrand of Vertex AI but a substantive evolution that layers major new governance, identity, security, and scaling capabilities (like cryptographic agent identity) on top of what Vertex already provided.Key New Capabilities✦ from: Gemini Enterprise Agent Platform is not a rebrand of Vertex AI but a substantive evolution that layers major new governance, identity, security, and scaling capabilities (like cryptographic agent identity) on top of what Vertex already provided.Macy's: Zero to Voice-Activated Agent Commerce✦ from: Macy's built a fully voice-activated, multi-agent commerce and service experience on Gemini Enterprise's customer experience platform in just four weeks, going from nothing to production.Not a Single Bot — An Orchestrated Agent Team✦ from: Macy's built a fully voice-activated, multi-agent commerce and service experience on Gemini Enterprise's customer experience platform in just four weeks, going from nothing to production.

Related episodes

Video · The Cognitive Revolution

An integrated, full-stack AI platform (chips to models to applications) eliminates the integration challenges that are enterprises' number one issue, cutting implementation and adoption costs, while chip-optimized model training (e.g., on TPUs) lowers token costs that get passed on to customers.

Gemini Enterprise Agent Platform is not a rebrand of Vertex AI but a substantive evolution that layers major new governance, identity, security, and scaling capabilities (like cryptographic agent identity) on top of what Vertex already provided.

Governance, security, and compliance capabilities are included as standard, built-in features of the Gemini Enterprise app rather than a separately priced add-on, because organizations shouldn't have to pay extra to adopt agents safely.

Openness means giving customers MCP endpoints for their data stores as a fallback to pre-built connectors, so they aren't blocked waiting on a vendor to build a specific connector before they can integrate their data into Gemini Enterprise.

Macy's built a fully voice-activated, multi-agent commerce and service experience on Gemini Enterprise's customer experience platform in just four weeks, going from nothing to production.

The main obstacle for enterprises adopting agents isn't the technology itself but choosing the right platform amid widespread 'agent washing,' where every vendor claims the same capabilities regardless of actual roadmap or delivered efficiency.

Fable's Back, AI Engineer Recap, & SambaNova

Software in the Age of Agents | The a16z Show

How I tricked Claude into leaking your deepest, darkest secrets