Meta launches Business Agent for enterprise operations

Meta Platforms unveiled the Meta Business Agent, an AI assistant meant to help businesses manage customer interactions and complete tasks on their behalf, at its WhatsApp-focused Conversations 2026 conference in London, Reuters reported. Per Meta, more than 1 million businesses already use earlier chatbot versions on WhatsApp and Messenger; the new agent will expand to Instagram and roll out globally. Naomi Gleit, Meta's head of product, called it "definitely an enterprise play," Reuters reported. Businesses can use the tool free initially, with paid subscription options planned; CNBC reported it will be bundled into a business tier of Meta One.

Meta describes "agentic" capabilities that go beyond rule-based automations, including answering business-specific questions, recommending products from a catalog, booking appointments, qualifying leads, and completing transactions. Meta also detailed a Business Agent Platform that connects to hundreds of non-Meta systems, citing integrations with Shopify, Zendesk, and Shopee, deployable across WhatsApp, Messenger, Instagram, and Meta Business Suite.

Reporting frames the launch as part of Meta's push to diversify revenue beyond advertising and places it alongside enterprise agent efforts from OpenAI, Anthropic, and Google. Companies with large consumer reach that add agentic features typically compete on two vectors: existing channel surface area and the depth of backend integrations that let agents take actions rather than only respond.

Agentic business assistants increasingly depend on production-grade connectors to commerce and ticketing systems; Meta's stated Shopify and Zendesk integrations fit that pattern. Bundling agents into existing messaging channels also shifts deployment considerations toward data flows between messaging, CRM, and the agent runtime, with authorization, audit trails, and human-escalation paths becoming central design questions.

• •Pricing and packaging: how free access transitions to paid tiers and how those map to usage limits and integration costs.
• •Enterprise controls: availability of logging, audit, role-based access, and data-residency features larger customers require.
• •Integration breadth and reliability: real-world quality of connectors and the agent's ability to complete end-to-end actions without manual intervention.