Google reframes Gemini as a cross-platform agent and cuts its top AI plan to $100

From assistant to agent

Google used its annual developer conference to recast Gemini less as a standalone chatbot and more as an intelligence layer that runs across its entire product line. Instead of asking users to open a separate app, the company described a model that can act inside Search, Android, Chrome, Workspace and YouTube, carrying out multi-step tasks on a person's behalf rather than simply answering questions.

The shift reflects a broader industry move from conversational tools toward autonomous agents that can plan, browse, fill forms and complete workflows with limited supervision. Google framed this as the next phase of its assistant strategy, positioning agentic capability as the main battleground for consumer AI.

An aggressive price move

Alongside the product changes, Google lowered the cost of its premium AI tier and raised usage limits, a combination widely read as an attempt to capture the emerging consumer agent market before rivals can scale comparable offerings. Cheaper access to a more capable model lowers the barrier for everyday users to adopt agent features at volume.

The pricing decision also pressures competitors who sell higher-priced premium plans, since it resets expectations for what a top-tier AI subscription should cost. For users, the immediate effect is more capability for less money; for the market, it signals that distribution and price, not just raw model quality, are becoming decisive.

Why it matters for developers and businesses

Embedding an agent across widely used apps could reshape how software is built and discovered, because tasks that once required visiting multiple services may increasingly be handled inside a single assistant. That has implications for app developers, marketers and anyone whose product depends on direct user visits.

It remains to be seen how reliably these agents perform real tasks outside controlled demos. Execution accuracy, safety and user trust will determine whether the agent model becomes a daily habit or a feature people try once and abandon.