Google Releases a Workplace AI Agent for Enterprises, but Pricing and Integrations Remain Open Questions

Google has released an agent for workplace tasks that uses a company's business context: the documents, data and processes accumulated inside the organization. AI Business reported the news on 9 October 2026. The release has already happened, but enterprise customers are left with two questions that have no clear answers: how much it will cost and how well the agent will coexist with tools that Google does not own. The available description contains no public pricing and no full list of integrations, so the product can only be judged on its architectural logic and the market context.

What we know so far

The core idea is straightforward. An ordinary chatbot answers general questions. A corporate agent sees a specific company's materials: internal reports, correspondence, knowledge bases. That lets it prepare summaries, search for information across scattered sources and carry out multi-step instructions. This is what people call knowledge work: analytics, document preparation, gathering data for decisions.

What the publication supports:

• Product class: an agent for work processes, not a consumer assistant.
• Key function: using corporate context when performing tasks.
• First risk area: total cost of ownership.
• Second risk area: integration with tools outside the Google ecosystem.
• Not disclosed in the available description: the pricing table, the list of connectors, the payment model (per user or per usage).

That last point matters. Without a payment model, the finance department cannot calculate payback, and without a list of integrations, the IT department cannot estimate how much extra work is needed.

Why money and integrations decide everything

Agents cost differently from an office suite subscription. A single user request to an agent often turns into a chain of dozens of model calls: search, reading documents, verification, assembling the result. If pricing is tied to usage, the budget becomes hard to predict. If it is tied to the number of seats, the company also pays for people who barely use the agent. The market already has a benchmark: Microsoft 365 Copilot costs 30 dollars per user per month, and buyers will compare any offer against that figure.

Integration is an even tougher problem. An agent's value equals the amount of data it can reach. Most large companies run a mixed landscape: email and documents in one place, tickets in Jira or ServiceNow, sales in Salesforce, finance in SAP, communication in Slack or Teams. Each of these links requires a connector, access-rights management and an audit of the agent's actions. If the agent reads Google files well but sees little else, it covers only part of the processes, and employees keep switching between systems by hand.

Security is a separate problem. An agent with broad access to corporate data must respect the permissions of each individual employee. A mistake in access separation turns a convenient search into a channel for leaking internal information.

Google's position is two-sided. Customers that live entirely in Google Workspace will find deployment easier. Companies with mixed infrastructure will have to convince themselves that the agent is worth the cost of connecting it. Competitors, from Microsoft to Salesforce, sell their own agent platforms and press on the same point: "we are already embedded in your processes."

How to approach a pilot

If you are considering this product, start with a narrow pilot of 20–50 employees and one or two scenarios where the result is easy to measure, such as report preparation or search across internal documentation. Ask the vendor up front for written answers to four questions: the payment model, usage limits, the list of supported external systems and the mechanism for inheriting access rights.

At the same time, establish a baseline: how many hours these tasks take now. Without it, three months from now you will not be able to tell real savings from impressions. And find out in advance what happens if one of your key systems turns out to be missing from the connector list: whether your IT department is ready to write the integration itself and what that would cost.