Apple Lets Poke Bring an AI Agent to iMessage Business Chat

Poke Gets Apple's Approval for iMessage

Poke, a consumer AI-agent startup built around text-based interaction, has become the first standalone AI agent approved to operate on Apple's Messages for Business platform.

That is a notable shift for Apple's business messaging system. Messages for Business has historically been used by companies such as airlines, retailers, hotels, and support teams to talk with their own customers inside iMessage. It supports a mix of automated flows and human agents, but it was not originally positioned as a distribution surface for independent AI assistants.

With Poke's approval, the platform is now being used for a different pattern: a user can interact with a general-purpose AI agent directly through iMessage, instead of opening a dedicated app or using a more technical agent interface.

What Poke Does

Poke launched earlier this year with a simple pitch: make AI agents feel as easy as texting. The product is aimed at everyday users rather than developers or power users who are comfortable with terminals, complex automation setups, or advanced agent frameworks.

The assistant can handle common personal tasks through messaging, including:

• planning the day and organizing tasks;
• managing calendar-related requests;
• tracking health and fitness information;
• controlling smart-home actions;
• helping edit photos;
• answering questions and coordinating follow-up steps through chat.

The company says the service has already handled roughly 100 million messages across its existing channels. Before the Apple approval, Poke was available through SMS, Telegram, and WhatsApp in selected markets. iMessage now becomes another supported surface.

Why Messages for Business Matters

Apple is not simply listing Poke as a normal consumer app here. Messages for Business is a structured communication channel inside iMessage, designed so users can contact a business without calling, installing a separate app, or navigating a support website.

For AI agents, that matters because messaging is one of the lowest-friction interfaces available. Users already understand how to make requests in a chat thread, wait for a reply, tap buttons, preview links, and continue the conversation later.

That makes iMessage potentially valuable agent infrastructure. Instead of asking users to learn a new dashboard, agent startups can meet them in the same place they already message friends, services, and support teams.

A New Distribution Cost for Agent Startups

The business model is just as interesting as the approval itself. Poke's parent company, The Interaction Company of California, will pay Apple on a per-user basis for access to the platform.

The exact price has not been disclosed, but the company describes Apple's fee as meaningfully lower than the cost of distributing third-party agents through Meta's messaging ecosystem after regulatory changes in Europe pushed Meta's fees higher.

That gives Apple a possible new revenue stream if AI agents become a major consumer category. It also creates a distribution cost that agent startups will need to model carefully. Messaging access may feel lightweight to users, but at scale it becomes platform rent.

For startups, the calculation is straightforward: if iMessage improves activation, retention, and trust enough, the per-user fee may be worth it. If not, agents may continue to prioritize lower-cost channels or direct app experiences.

Apple's Approval Requirements

Getting onto Messages for Business was not instant. Poke had to satisfy Apple's requirements around support, disclosure, and interface behavior.

The approval process included several practical checks:

• Poke had to show that live human support was available when needed;
• the AI system had to be clearly presented as an AI agent rather than disguised as a person;
• the company submitted supporting documentation from its messaging providers;
• the iMessage experience had to follow Apple's interface guidelines;
• links, buttons, and interaction elements had to be adapted to Apple's expected style.

One visible change is link handling. Poke's iMessage version needs to use Apple-style link previews rather than simply dropping inline links into the conversation. Buttons and other interface elements also need to match the Messages for Business design system.

That means future AI-agent companies should expect the approval process to take time. Poke spent months adjusting its experience to fit Apple's requirements, and similar companies will likely face the same kind of review.

Timing Ahead of WWDC

The approval lands just before Apple's Worldwide Developers Conference, where Apple is expected to show more of its AI strategy, including updates around Siri, Apple Intelligence, and tools for app developers.

This does not necessarily mean Apple is opening a full AI-agent App Store. Messages for Business is not the same thing as a consumer app marketplace. Still, the Poke approval suggests Apple is willing to let AI agents operate inside controlled Apple-owned surfaces if they meet the company's rules around trust, support, and user experience.

That fits Apple's usual platform posture: open the door, but keep the interaction model tightly governed.

Why This Is Bigger Than One Chatbot

Poke's iMessage rollout is another sign that consumer AI agents are moving from experiments into distribution battles. The core question is no longer only whether agents can reason, call tools, or complete tasks. It is also where users will actually talk to them.

The answer may be less about standalone apps and more about existing communication channels. If users can ask an agent to schedule, plan, buy, edit, summarize, or control devices from the same messaging thread they already use every day, adoption becomes easier.

That mirrors a broader shift across agentic products. Robinhood is experimenting with AI agents that can act inside financial accounts, while other companies are building agents for commerce, productivity, and personal operations. For more on that trend, see our coverage of Robinhood's AI-agent trading accounts and our guide on how to build an AI agent.

The big challenge is trust. Messaging makes agents more accessible, but it also makes them feel more personal. Apple's requirements around live support, disclosure, and interface consistency show how seriously platforms are treating that trust layer.

Poke now has the first-mover advantage on iMessage. If the experience works, other consumer AI-agent companies will likely try to follow — and Apple may have found a new way to monetize the agent layer without needing to release a standalone agent marketplace first.