Google brings agentic AI to Gemini, starting with businesses

Google unveiled its advanced AI system, Gemini, at a recent Cloud event, marking a significant leap in the realm of agentic AI. This unified agent, accessible through a single interface, transcends the role of a mere conversational assistant, now equipped to perform tasks initiated by users. The innovation aligns with the broader industry trend towards AI systems capable of autonomous task execution, data manipulation, and complex operations like code generation, scheduling, and meetings booking.

The announcement comes in the wake of consumer-facing agents like Meta's Muse and messaging-based agents such as Instinct, reflecting a growing interest in AI that operates beyond traditional boundaries. Moreover, the debut of ChatGPT's Dots further underscores the evolution of AI capabilities.

Google's CEO, Sundar Pichai, highlighted the massive potential of Gemini, noting that it currently boasts over 1 billion monthly active users. Furthermore, a substantial portion of Fortune 100 businesses, approximately 90%, already utilize Gemini Enterprise for work-related tasks. Given these impressive figures, Google prioritized the rollout within the enterprise sector before extending the functionality to consumers.

Security, scalability, and performance were identified as key challenges for deploying powerful AI agents such as Gemini. To address these concerns, the company initially aimed to refine the agent's security, adaptability, and operational efficiency within corporate environments. Thomas Kurian, CEO of Google Cloud, emphasized the agent's ability to receive objectives instead of simple instructions, enabling it to strategically plan tasks, leverage custom skills, and interact with businesses' internal systems.

The agent is designed to automatically select the most suitable model for task completion, but users retain the flexibility to switch models, including third-party options like Anthropic's Claude models. The system's adaptability is further enhanced by the inclusion of attachments such as files, folders, and specific workstream projects.

Moreover, the agent can securely connect to a wide array of business data and systems, including Google Workspace, Microsoft 365, Slack, Jira, Confluence, Git, BigQuery, Databricks, Postgres, and Snowflake, among others.

A unique feature of Gemini is its "tasks inbox" interface, offering users visibility into the AI's decision-making process, delegation of tasks, and progress tracking. Gemini is given its own Workspace account, complete with an email address and contextual awareness of the company's structure, schedules, approvals, and more. Users can interact with the agent through various channels, including tagging, emailing, group chats, and calendar integrations.

The agent will be accessible across multiple platforms, including mobile devices, desktops, and command-line interfaces, as well as within popular enterprise applications like Workspace, Microsoft 365, ServiceNow, and Slack. Early adopters of Gemini Enterprise include sportswear brand On, e-commerce platform Shopify, and payment processor PayPal.

Notably, large corporations such as BNP Paribas, Bradesco, Merck, Orange Spain, Santee Cooper, SOMPO, Ulta Beauty, and Wesfarmers are among the early users. Google also introduced flexible spending options, including multi-model orchestration, smart routing, and real-time spend caps, to help businesses manage the costs associated with their enterprise AI usage.