For most of the past decade, Meta's business model could be summed up in a single word: advertising. The company that owns Facebook, Instagram, WhatsApp and Messenger built one of the most profitable machines in technology by selling attention. Now it is trying to sell something else entirely — labor. With the global launch of the Meta Business Agent and the broader Business Agent Platform, Meta is making its most concrete attempt yet to turn its messaging empire into an enterprise software business, and to give companies of every size an always-on AI employee that lives inside the apps their customers already use.
What Happened
Meta announced the worldwide availability of the Meta Business Agent at its Conversations conference, ending nearly two years of limited testing that had been confined to a handful of markets including India, Mexico and Brazil. The agent can now be switched on by any business on WhatsApp, Instagram and Messenger, and the company says it takes only minutes to configure one. The pitch is straightforward: instead of a customer waiting on hold or filling out a contact form, they message a business the way they would message a friend, and an AI agent answers around the clock.
The capabilities go well beyond canned replies. According to Meta, a Business Agent can field customer inquiries, recommend products pulled directly from a company's catalog, schedule appointments, qualify sales leads and even help complete a transaction without a human ever stepping in. Alongside the consumer-facing agent, Meta introduced the Business Agent Platform — connective tissue that links agents to hundreds of third-party systems such as Shopify, Zendesk and Shopee, plus enterprise-grade controls, measurement tools and built-in guardrails aimed at larger organizations. Crucially, those integrations let an agent take actions on a company's behalf rather than simply respond to messages. The service is free for now, with paid subscription tiers and pay-per-token pricing promised in the coming months.
Why It Matters
The scale behind the launch is what makes it significant. Meta says more than one million businesses already use an AI-powered agent on WhatsApp and Messenger, and that more than one billion conversations between businesses and customers take place every day across its messaging apps. That existing footprint gives Meta a distribution advantage that pure-play enterprise startups simply cannot match — the customers, the conversations and the commerce are already flowing through its pipes.
It also represents a deliberate hedge. Advertising still accounts for the overwhelming majority of Meta's revenue, and a business that sells AI agents on a usage basis is a fundamentally different proposition from one that auctions ad impressions. By charging companies for the work an agent performs — rather than only for the eyeballs it reaches — Meta is positioning messaging as a second commercial engine. For small merchants in particular, the appeal is obvious: an AI agent that handles inquiries and sales at all hours lowers the cost of running a storefront and competes with the round-the-clock service that only large companies could previously afford.
Reaction
The response from the business community has been a mix of enthusiasm and caution. Small and mid-sized merchants, who were the focus of Meta's pilot programs, have been the most receptive, drawn by the promise of automating repetitive customer service without hiring a support team. Meta has gone as far as to claim that the tools can help some companies multiply their output dramatically, a figure that analysts have treated with healthy skepticism. Developers, meanwhile, have focused on the Business Agent Platform's integrations, since the ability to wire an agent into Shopify orders or a Zendesk ticket queue is what turns a chatbot into something closer to an operational system.
Not everyone is convinced. Some enterprise buyers have raised familiar concerns about handing customer interactions to an autonomous system — accuracy, brand voice, data governance and what happens when an agent gets something wrong on a transaction. Meta's emphasis on guardrails and measurement tools is a direct answer to those worries, but trust in agentic systems that act on a business's behalf will be earned slowly, deal by deal, rather than granted on launch day.
What's Next
The most important variable is pricing. Meta has said the paid tiers will arrive in the coming months on a subscription model with pay-per-token costs, and the details will determine whether the Business Agent becomes a genuine revenue line or a feature that mostly serves to keep businesses engaged with the ad platform. The company also faces a crowded field: Salesforce, ServiceNow and a wave of startups are all racing to sell enterprise AI agents, each arguing that its system can sense, decide and act on a company's behalf.
Expect Meta to lean on its scale. The roadmap points toward deeper commerce features, more third-party integrations and tighter ties between agents and the catalogs, payments and advertising tools that already run on its apps. If the company can convince businesses that an agent on WhatsApp is not just a support tool but a salesperson, the line between a customer conversation and a completed purchase could keep getting shorter.
Closing Thoughts
Meta's Business Agent is less a single product than a statement of direction. The company is betting that the future of business software is conversational, that the interface is the chat thread people already keep open all day, and that the most valuable thing AI can do for a company is not write a poem but close a sale. Whether that bet pays off will depend on execution, pricing and trust — but the ambition is unmistakable. Meta is no longer content to be the place where businesses buy attention; it wants to be the place where businesses do the work.
.jpg){kind=link}
For the wider industry, the launch is another sign that 2026 is the year agentic AI moved from demo to deployment. The interesting question is no longer whether AI agents can hold a conversation, but whether they can be trusted to act — and which platform will own the moment a customer's message turns into a transaction. Meta has just placed a very large chip on that table.
한글 요약
메타가 자사 콘퍼런스에서 '메타 비즈니스 에이전트'를 전 세계에 정식 출시하며 광고 중심이던 사업 구조를 확장하려는 가장 구체적인 행보에 나섰다. 인도·멕시코·브라질 등 일부 시장에서 약 2년간 시험 운영을 거친 이 AI 에이전트는 이제 와츠앱·인스타그램·메신저에서 모든 규모의 기업이 몇 분 만에 설정해 24시간 고객 응대에 활용할 수 있다. 단순 자동응답을 넘어 카탈로그 기반 상품 추천, 예약, 리드 선별, 결제까지 사람 개입 없이 처리하도록 설계됐다.
함께 공개된 '비즈니스 에이전트 플랫폼'은 쇼피파이·젠데스크·쇼피 등 수백 개 외부 시스템과 연동돼 에이전트가 단순 응답이 아니라 실제 업무를 대신 수행하도록 한다. 메타는 이미 100만 개 이상의 기업이 와츠앱·메신저에서 AI 에이전트를 쓰고 있고 하루 10억 건 이상의 기업-고객 대화가 오간다고 밝혔다. 현재는 무료지만, 향후 구독형과 토큰 사용량 기반 유료 요금제가 도입될 예정이다.
업계는 메타의 압도적 사용자 규모를 강점으로, 정확성·데이터 거버넌스·신뢰 문제를 과제로 본다. 세일즈포스·서비스나우 등과의 경쟁이 치열한 가운데, 핵심 관전 포인트는 요금제 설계와 '대화가 곧 거래로 이어지는' 메타식 통합 전략의 성패다. 이번 출시는 2026년이 에이전트형 AI가 시연을 넘어 실제 배치 단계로 진입한 해임을 다시 한번 보여준다.