Meta has flipped the switch on a worldwide rollout. The company announced Wednesday that its customer support bot, now called Meta Business Agent, is available globally inside WhatsApp, according to TechCrunch AI. This is the payoff after nearly two years of testing AI agents in WhatsApp Business in markets like India and Mexico.
The play here is bigger than a chatbot. Meta wants to turn WhatsApp from a messaging layer into actual workflow software for small and medium businesses. For a billion-plus app that companies already lean on to talk to customers, that’s a serious shift in ambition.
Here’s what the Business Agent actually does, what’s still cooking, and what it’ll cost you.
What the agent can do today
TechCrunch AI reports the bot handles the core customer-facing jobs a small business juggles every day:
- Answer customer questions. It fields incoming queries directly in chat, so a human isn’t tied to the inbox around the clock.
- Recommend products. The agent can surface relevant items, nudging a casual conversation toward a sale.
- Book appointments. Scheduling happens inside the same thread, no extra app or back-and-forth.
- Qualify sales leads. It sorts the serious buyers from the browsers before a person ever steps in.
- Reroute to a human when needed. If the question is too complex, the agent hands it off rather than guessing.
Meta is also extending the bot beyond WhatsApp into Instagram DMs, which widens its reach to the businesses that live on Instagram’s audience.
What’s still in testing
Meta isn’t done. The company is trialing a feature that gives businesses daily briefings of conversations that happened overnight, plus insights drawn from those chats. That’s being tested with select accounts across WhatsApp Business, Instagram Pro, Messenger, and Meta Business Suite.
A longer roadmap is taking shape too. Meta says it’s working to let the agent:
- Run market research
- Highlight product features
- Manage users’ calendars
- Connect with tools to pull competitive insights
- Surface businesses when someone searches for one, and share contact details in chat
What stands out to me is the enterprise angle. Meta is building a platform for larger companies to create custom agents that plug into systems like Shopify, Zendesk, and Shopee. That’s where this stops being a small-business convenience and starts looking like a real commercial product.
What it costs
None of this is charity. Meta plans to charge businesses by folding the agent into some tiers of its WhatsApp Business Premium subscription. Larger businesses will pay based on how many tokens they use, meaning the heavier the usage, the bigger the bill.
That pricing detail matters more than it might seem. As TechCrunch AI notes, WhatsApp has leaned on businesses paying for messaging and click-to-WhatsApp ads. An AI agent baked into a paid tier gives Meta a fresh, usage-based revenue line on top of an audience it already monetizes.
Why it matters
Meta is meeting small businesses where they already work. Most SMBs don’t run dedicated support software, and a lot of them already handle customers over WhatsApp out of habit. Dropping an AI agent into that existing behavior removes the friction of adopting a whole new tool.
The competitive context is worth naming. Standalone customer-support AI and chatbot platforms have to win businesses over from scratch. Meta starts with distribution most of them would envy, the contacts, the chat history, and the daily usage are all already there. The token-based pricing for big accounts mirrors how AI infrastructure broadly gets billed now, so enterprises won’t find the model unfamiliar.
The caveats are real, though. The most interesting capabilities, overnight briefings, calendar management, competitive research, and the custom enterprise agents, are still in testing or on the roadmap, not shipped. And businesses that lean on this will be wiring another core workflow directly into Meta’s ecosystem, with the lock-in that comes with it.
For now, the launch is live and global. The features that decide whether this becomes essential software, rather than a nice-to-have, are the ones Meta hasn’t finished building yet. You can find the full details at the original TechCrunch AI report.