Palo Alto startup Poke is betting that the future of artificial intelligence will look less like a dashboard full of knobs and more like a simple text thread on your phone. By turning powerful AI agents into contacts you can message on iMessage, SMS, Telegram, and, in some markets, WhatsApp, the company says it can make advanced automation feel as natural as “texting a friend to get something done.”
Poke, built by The Interaction Company in Palo Alto, California, lets users access a personal AI assistant entirely through familiar messaging apps, with no new app downloads or complex onboarding required. Once connected, the AI agent can handle tasks like daily planning, calendar management, health tracking, smart‑home control, email triage, and more all triggered by simple text instructions instead of forms or settings menus.
The service officially launched to the public in March 2026, after an early rollout that began with iMessage and SMS and later expanded to Telegram and WhatsApp in select regions. Rather than asking users to learn a new interface, Poke “shows up as a contact” inside your existing messaging apps, promising “zero friction” access to an AI that is always one text away.
Poke’s launch lands at a moment when so‑called “agentic AI” AI systems that can not only respond to prompts but also act on a user’s behalf are moving from research labs and developer tools into mainstream consumer products. Instead of requiring users to wire up complex workflows or integrations themselves, Poke connects behind the scenes to email, calendar, files, and other services so that everyday instructions like “move my 3 p.m. meeting to tomorrow” or “find a cheaper flight for Friday” can be carried out with a single message.
Poke describes its philosophy simply: your phone is already the hub for your digital life, so the smartest place for an AI agent to live is inside the messaging apps you already use. In a write‑up on the service, one reviewer noted that when they first texted Poke, “its responses to my initial text were so natural and human‑like that I immediately knew [the team] was onto something.”
Getting started with Poke begins with sending a text to the service and connecting at least one email account, which the agent uses to understand your schedule, responsibilities, and communication patterns. After that, users can type instructions in plain language with no coding or technical setup required and Poke’s AI translates those requests into automations that run across connected services.
Today, the assistant can manage a growing list of routine tasks: it can handle daily planning, send medication reminders, track health and fitness goals, control compatible smart‑home devices, surface important emails, or summarize the previous night’s sports scores all via short text exchanges. Users can ask Poke to “alert me to emails from my boss,” “remind me at 8 a.m. if I need an umbrella,” or “catch me up on the day’s news,” and the agent will proactively message them with timely updates.
One of the more ambitious features is the ability for users to describe custom automations in natural language and then share those automations with friends, effectively turning ordinary text messages into a distribution channel for reusable AI workflows. Poke’s creators argue that this “write it in plain text” approach lowers the barrier to building and using AI agents in the same way that early no‑code tools lowered the barrier to building apps and websites.
Beyond its technical capabilities, Poke is drawing attention for its conversational style. In early user accounts, the agent is described as informal, witty, and even a little confrontational at times, with one reviewer writing that “Poke’s informal lowercase sentences and casual verbiage” make it “feel so much more human” than traditional chatbot interfaces.
In one anecdote, the AI greets new users with a mix of helpfulness and teasing, then quickly pivots into setting up accounts and asking for permissions to access email and calendars. The experience, according to that account, “immediately feels like you’re shooting the breeze with a close friend” rather than configuring a piece of enterprise software.
That human‑like tone is intentional. Poke’s team is betting that personality can help ease users into trusting an AI agent with sensitive tasks like inbox management, travel bookings, and personal reminders, while still maintaining enough transparency and control to keep people comfortable.
Though still young, Poke has already attracted significant investor attention. According to recent reports, the company raised an initial seed round of around 15 million dollars in 2025 and added a further 10 million dollars in early 2026, bringing its valuation to roughly 300 million dollars around the time of its public launch this March.
Backers include well‑known firms such as Spark Capital and General Catalyst, whose involvement signals confidence that messaging‑native AI agents could represent a new consumer platform rather than just another productivity app. Following the March rollout, Poke has reportedly seen a “10x increase in signups” in April, suggesting that the concept is resonating with early adopters beyond the usual tech‑savvy crowd.
Industry observers note that Poke’s approach stands in contrast to enterprise‑focused agentic systems being developed by large AI providers, which often prioritize complex integrations and business workflows over everyday convenience. By targeting consumers first and framing the product as a simple contact you can text for help, Poke is positioning itself as a gateway to AI agents for users who might never touch developer tools or automation platforms.
Poke’s arrival also highlights a broader shift underway in AI: from conversational models that respond to questions to agents that can perceive, reason, and act with a degree of autonomy. Researchers and industry experts describe “agentic AI” as the next phase of generative AI, where systems connect to external tools and services, carry out multi‑step tasks, and adapt over time with minimal supervision.
In this context, Poke can be seen as a consumer‑friendly wrapper around emerging agent capabilities. Instead of exposing users to APIs, workflows, or configuration screens, it hides the complexity behind a text thread while still giving them the feeling of directing a capable digital assistant that understands context and can take initiative.
For all its promise, Poke still faces critical questions common to AI assistants that handle sensitive personal data. Connecting to inboxes, calendars, health information, and smart‑home devices raises issues around privacy, security, and data governance that the company will need to address clearly if it hopes to win mainstream trust.
There is also the challenge of reliability. Agentic systems that can autonomously act on users’ behalf must not only understand instructions but also execute them accurately and safely, especially when dealing with financial decisions, travel bookings, or medical reminders. Any missteps could quickly erode user confidence in a product that is supposed to quietly run much of their digital life in the background.
Finally, Poke will have to prove that its messaging‑first approach scales beyond early enthusiasts. As more platforms and competitors move into the AI agent space, the company will be under pressure to continue innovating on both capabilities and user experience without losing the simplicity that makes “just texting” so appealing in the first place.
For now, Poke’s pitch is straightforward: instead of downloading yet another app or learning a new automation tool, you just send a text and let an AI agent handle the rest. With strong backing, rapid early growth, and a product that meets users where they already are in their messaging threads the startup is emerging as one of the more closely watched experiments in bringing AI agents to everyday life.
If its early promise holds, the most powerful software on your phone may soon look less like an app icon and more like a chat bubble waiting for your next message.
Comments