Google and OpenAI both ship agent platforms — same week, same bet

Within 48 hours, Google launched the Gemini Enterprise Agent Platform and OpenAI shipped Workspace agents plus GPT-5.5. The synchronized timing is the real signal.

In a 48-hour window between April 22 and April 23, 2026, Google and OpenAI both unveiled enterprise agent platforms, and OpenAI shipped a new flagship model explicitly framed around agent workflows. The coincidence is the story: the industry has collectively decided that agents — not chat, not copilots — are the interface layer enterprise buyers will pay for.

Google used Google Cloud Next to announce the Gemini Enterprise Agent Platform, positioned as the evolution of Vertex AI. The pitch focuses on the unglamorous middle of the stack: integration, security, governance, and optimization. The Model Garden gives access to 200+ models, including Gemini 3.1 Pro, Gemini 3.1 Flash Image, and Gemma 4 (all as of April 2026). Google is betting enterprises want a place to build agents, not a new model to evaluate.

OpenAI ran a parallel play. Workspace agents arrived in research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans, with integrations that let agents update Linear issues, post in Slack, pull context from docs and email, and route work across teams. A day later, OpenAI announced GPT-5.5, rolling out to Plus, Pro, Business, and Enterprise users in ChatGPT and Codex as of April 23, 2026. The framing is notable: OpenAI describes the model as built to understand complex goals, use tools, check its work, and carry tasks through to completion — agent-first language, not chatbot language. OpenAI claims GPT-5.5 matches GPT-5.4 per-token latency while using significantly fewer tokens on Codex tasks (OpenAI's own measurement, April 2026 — independent benchmarks aren't out yet).

Researcher Jason Liu captured the technical arc in a single line: "First, we had to solve reasoning, and then we solve agents. We didn't try to skip the reasoning part." That sequencing explains why the agent wave is arriving now rather than two years ago — the reasoning substrate had to land first.

For builders, model choice is becoming less load-bearing than platform choice; integration depth and governance primitives are where the lock-in lives. For buyers, the question shifts from "which model is best" to "which platform do we trust with production workflows and approved actions." For anyone watching vendor strategy, the synchronized launches suggest the window to differentiate on model capability alone is closing fast.

The bottom line: If agents are the new enterprise interface, pick the platform by its integration and governance story, not by last week's benchmark leader.

Sources