On client call after client call, I hear the same confusion:
“Wait, is Vertex AI gone?”
“Is Gemini Enterprise and the Agent Platform the same thing?”
“So do I build my agent in Gemini Enterprise now, or somewhere else?”
If you are confused by Google’s AI rebrand, you are not alone. The good news: almost nothing about how you build has actually changed. What changed is the map, not the territory.
Here is the version I now give every team that asks.
What actually happened
At Cloud Next 2026, Google unified its AI portfolio under one brand. Nothing was removed. Things were renamed and consolidated so that the whole stack tells one story: the agentic era.
The renames that matter:
| Before | After |
|---|---|
| Agentspace | Gemini Enterprise (the app) |
| Vertex AI | Gemini Enterprise Agent Platform |
| Gemini Code Assist | Still here, now part of Gemini Enterprise editions |
Google describes the Agent Platform as the evolution of Vertex AI: future Vertex services and roadmap items now ship through the Agent Platform instead of as a standalone product.
And the part everyone worries about first: your existing Vertex AI SDKs, APIs and billing keep working. No breaking changes. The endpoints you shipped on last quarter still answer today.
Two pillars, one brand
Once you stop reading it as “everything moved” and start reading it as “two clear pillars,” it gets simple.
Gemini Enterprise
The AI workspace for employees: enterprise search, chat with your own data, deep research and no-code assistants. Where people use AI across a company.
Best for business + knowledge workers
Gemini Enterprise Agent Platform
The Agent Development Kit (ADK) for code-first agents, Agent Studio for low-code, 200+ models via Model Garden, a managed runtime in Agent Engine, plus registry, identity and governance. Where teams build AI.
Best for developers + platform teams
If you remember one line, remember this:
Use Gemini Enterprise to consume AI. Use the Agent Platform to build AI.
That single sentence answers most of the questions we get.
So where do I build my agent?
You build on the Agent Platform, and you can surface that agent inside Gemini Enterprise for your employees. The two pillars are not competitors. One is the workbench, the other is the storefront.
A typical build flow looks like this:
- Define the agent with ADK: tools, instructions, and the model you want, in code you can test and version.
- Run it on Agent Engine: the managed runtime, so you are not babysitting infrastructure.
- Register it in the Agent Registry, with identity and governance attached.
- Expose it in Gemini Enterprise, so the people who need it can just use it, no terminal required.
That last step is the “aha” for most teams: the thing you build on the Agent Platform becomes a first-class assistant your colleagues open like any other app.
Registering an ADK agent so it shows up inside Gemini Enterprise is exactly the kind of step we walk teams through. If you have shipped on Vertex AI before, it will feel familiar fast.
What this means in practice
Rebrands are noisy, but the signal underneath this one is clear: Google is betting the stack on agents, and it wants one coherent place to build them and one coherent place to use them.
For most teams, the move is not a migration. It is a vocabulary update plus a decision:
- If you mostly want your team to use AI on top of your own data, start in Gemini Enterprise.
- If you are building agents that do real work in production, you live in the Agent Platform, with ADK at the center.
That is the whole map. No, Vertex AI did not disappear. It grew up and got a new name.
This is the kind of question we field constantly: not “what was announced,” but “what do we actually do with it.” Clarity like this is half of what we do at Oloodi. The other half is building the agents.
Boris-Wilfried Nyasse, Founder & Google Developer Expert at Oloodi.