Gemini at Work 2026: Google Expands AI Capabilities From Individuals to Teams

Google has expanded the role of its Gemini workplace AI with a new agent designed to work not only for individual employees but also on behalf of teams and specific roles within an organisation. The company announced the development at Gemini at Work 2026, positioning the agent as an AI system capable of handling knowledge work, answering questions, creating media, writing and running code, and carrying out tasks based on an overall objective rather than requiring users to provide every individual step.

Google said the Gemini agent can function as a “team member” working for a group, such as a project manager, or perform the responsibilities associated with a particular organisational role, such as an analyst in a finance department. The system can also take on assignments that would normally require coordination among multiple employees. For example, if a manager asks an employee for an updated project report in the form of a slide deck, Gemini can identify the task and offer the employee a one-click option to have the agent complete it.

The agent is also designed to assist with workplace coordination and communication. Google said an employee could ask Gemini to organise a meeting involving a regular group of regional event leads without providing their individual names or email addresses. The agent can identify the relevant people through membership of a workplace chat space and previous conversations, check calendars and begin an email discussion to determine a suitable meeting time. It can also account for situations involving people from outside the organisation.

Also Read: Pakistan Jets Reportedly Strike Houthis in Yemen Under Mecca Pact

Google said organisations can create “coworker agents” with defined responsibilities and separate identities. Such agents can have company email addresses and persistent storage, while their access to information remains limited to data provided by the user or team. Gemini can also create temporary sub-agents to handle different parts of a larger assignment. Because the agent operates in the cloud, tasks that take hours or days can continue even after an employee closes their laptop. It can be accessed across web, Android, iOS, Windows and Mac, as well as through workplace services including Google Workspace, Microsoft 365 and Slack.

The Gemini agent can also connect with a range of business software and databases, including Salesforce, ServiceNow, Jira, Git, BigQuery, Databricks, Postgres and Snowflake. Companies can create their own tools and reusable skills for the agent, expanding the types of work it can perform. The development comes as businesses increasingly adopt AI for workplace tasks, although research from Anthropic indicates that AI exposure remains uneven across professions. Its analysis of real-world Claude usage found relatively low exposure in occupations involving physical activity, real-world interaction and responsibilities that current AI systems cannot easily perform.

Anthropic's research identified areas such as agriculture, construction, installation and repair, transportation, personal care and food services as having lower observed AI exposure. Agricultural workers, for instance, may need to operate machinery or perform physical tasks, while legal professionals representing clients in court still require physical presence.

In contrast, computer programmers showed AI involvement in around 75 per cent of tasks, while customer service representatives and data entry workers were also among the more exposed occupations. The research also highlighted a gap between potential and actual AI adoption: in computer and mathematics occupations, AI could theoretically assist with nearly 94 per cent of tasks, but observed Claude usage covered about 33 per cent.

Also Read: Over 5,000 Passengers Miss Trains Amid INDIA Bloc Protest in Delhi