Poke Joins Apple's Messages App as AI Agent

9to5Mac reports that Poke, a third-party AI agent, is now available inside Apple's Messages app on iPhone, letting users chat with it from iMessage and have it act on requests, in what 9to5Mac calls the first third-party AI agent to run inside Messages. 9to5Mac attributes the integration to Apple's Messages for Business (Business Chat) pathway. Poke, made by The Interaction Company of California and co-founded by Marvin von Hagen and Felix Schlegel, launched publicly in March and already works over iMessage, SMS, and Telegram, handling tasks like calendar, email, reminders, and smart-home control, according to TechCrunch. Apple's Messages for Business platform already supports AI-powered bots and virtual agents for companies, and TechCrunch reported Poke previously reached iMessage by leveraging Linq, a service that lets AI assistants operate inside messaging apps. 9to5Mac and some users reported reply delays for the new Messages integration shortly after launch.
What happened
9to5Mac reported on June 4 that Poke, a third-party AI agent, is now available inside Apple's Messages app on iPhone, letting users message it from iMessage and have it take action on their requests. 9to5Mac characterized this as the first third-party AI agent to operate inside Messages and attributed the integration to Apple's Messages for Business (Business Chat) pathway. 9to5Mac and some users on social media reported that the service returned slow or failed replies for some people shortly after launch.
What Poke is
Poke is an AI agent from The Interaction Company of California, a Palo Alto startup co-founded by Marvin von Hagen and Felix Schlegel, TechCrunch reported in April. It launched publicly in March and is designed to be used entirely through messaging, iMessage, SMS, and Telegram, with no separate app. Per TechCrunch, Poke can manage calendars, monitor email, track health and fitness, control smart-home devices, and run user-written automations called "recipes" that connect to services such as Gmail, Google Calendar, Notion, and Linear. TechCrunch reported that Poke routes each task to whichever AI model fits best, whether from a major provider or an open-source model. The company is backed by General Catalyst and Spark Capital.
How agents reach Messages
Apple's Messages for Business platform already lets companies run customer-facing conversations in Messages and supports AI-powered bots and virtual agents alongside human escalation, according to Apple's description of the service. Separately, TechCrunch reported that Poke previously reached iMessage by leveraging Linq, a service that enables AI assistants to live within messaging apps; 9to5Mac noted in March that such conversational AI services have historically relied on Mac mini server farms to send iMessages.
Why it matters
Embedding an agent directly in a default messenger removes the need to install a separate app or open a browser, which can lower the barrier to trying conversational AI. Comparable integrations have historically boosted early engagement while surfacing operational challenges, scaling inference, formatting responses inside a chat UI, and meeting platform content rules. The reply delays reported at launch by 9to5Mac illustrate a common friction point when a service routes traffic through an intermediary platform channel.
What to watch
Whether Apple approves additional third-party agents through the same route, whether it publishes formal developer guidance for assistant integrations in Messages, and whether reliability or moderation limits reshape how these agents operate. Signals to monitor include approval volume, documented Apple guidance, and uptime and latency reports from operators.
Scoring Rationale
The story marks a notable platform-level integration: a first third-party AI agent inside iMessage increases developer and practitioner attention. Impact is moderate because it currently relies on a Business Chat workaround and shows early reliability issues.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems