Two weeks ago at Google Cloud Next ’26 in Las Vegas, Google did something the enterprise AI industry has been dancing around for the better part of two years. This means making agent-based AI governance a native product feature rather than an afterthought.
The headline announcement was the Gemini Enterprise Agent Platform, pitched as the successor to Vertex AI and described by Google as a comprehensive platform for building, scaling, managing, and optimizing agents. What made it notable wasn’t the model access or TPU upgrades, although those are important.
This was its underlying architecture. Every agent built on the platform gets a unique cryptographic ID for traceability and auditing, and the agent gateway handles monitoring interactions between agents and corporate data. In other words, governance comes with the product.
This design choice is a direct response to the issues that are quietly undermining AI adoption across the enterprise.
The governance gap that no one wants to talk about
A survey of 1,879 IT leaders published by OutSystems in April makes the numbers clear. 97% of organizations are already considering an agent AI strategy, and 49% describe their capabilities as advanced or expert. However, only 36% have a centralized approach to agentic AI governance, and only 12% use a centralized platform to maintain control over AI sprawl.
That’s an 85-point gap between confidence and actual control, and it’s not improving fast enough. Gartner’s 2026 Hype Cycle for Agentic AI expresses the same tension in a different way. Only 17% of organizations have actually deployed AI agents to date, but more than 60% expect to do so within two years, the most aggressive adoption curve Gartner has recorded for an emerging technology in the history of the study.
The hype cycle has placed agent AI right at the pinnacle of expectations, with governance, security, and cost management capabilities still maturing far behind the intended adoption. The reality of production is quite grim. Multiple independent analyzes estimate that the percentage of agent AI pilots that reach full-scale production scale is between 11% and 14%. The remaining 86% to 89% are stuck, quietly shelved, or never moved beyond a proof of concept.
Broken governance and complexity of integration are consistently cited as the main causes, rather than technical flaws in the model itself.
What Google is really betting on
Google’s message at Cloud Next ’26 was less about model capabilities and more about control plane ownership. Bain & Company’s post-event analysis found that Google is repositioning away from model access toward a complete agent-based enterprise platform where context, identity, and security are at the center of the architecture rather than at the edges.
The strategic logic is consistent. All three major cloud providers announced their agent registries for the first time in April 2026. This indicates that industry-wide governance tools are still in their infancy. Google’s move is the most comprehensive response yet, but it also has special implications for companies evaluating the platform. That means deeper integration with Google’s stack is part of the deal.
The tension between the true governance capabilities provided and the platform commitment required to access them is what enterprise architects are currently grappling with. Agent systems grow identities and privileges at a pace that traditional human-centric identity and access management models could never handle.
Once agents start operating system-wide, governance questions move from which models are authorized to what specific agents can run through which identities, against which tools, and with what audit trails.
Google’s cryptographic agent identity and gateway architecture is a direct answer to that question. Whether companies are ready to hand over that level of operational centricity to Google is another story.
Agent cleaning makes this difficult
There are complex issues that governance discussions tend to avoid. That is, most of what is currently being sold as agentic AI is not agentic AI. Deloitte research on enterprise AI trends points out that many so-called agent initiatives are actually hidden automation use cases. It is a traditional workflow tool with a conversational interface and works based on predefined rules rather than reasoning towards a goal.
This distinction is important because governance frameworks designed for truly autonomous agents do not map cleanly to scripted automation, and vice versa. Companies that confuse the two end up adopting governance structures that are either too restrictive for real agents or too permissive for weak automation masquerading as intelligence.
Gartner estimates that more than 40% of agent AI projects could be canceled by 2027, citing value uncertainty and weak governance as key reasons. You should be able to concentrate on that image. Companies that invest in governance architecture today (audit trails, escalation paths, limited autonomy, agent-level identity) are laying the foundation for determining whether their agent deployments will survive contact with the production environment.
The launch of Google’s Cloud Next platform is, at the very least, a mandatory feature. Tools for managed agent systems are now available at scale from major providers. What remains is the more difficult organizational work of deciding what agents are actually authorized to do, who will be held accountable if they make a mistake, and whether the platform that brings it all together is something they are ready to build on.
See also: SAP: How enterprise AI governance ensures profit margins
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events such as Cyber Security & Cloud Expo. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
