Is Poke the approachable AI assistant ordinary people have been waiting for? That’s the promise driving a new startup’s creation—a digital helper accessible through the messaging apps folks already rely on: iMessage, SMS, Telegram, and in select regions, WhatsApp.
Poke quietly arrived on the scene this March, designed as a personal, proactive agent capable of handling daily tasks straight from your chat window. Forget installing new software or navigating unfamiliar dashboards—just message Poke to organize your life. Need a hand with your calendar, fitness routines, morning reminders, smart home gadgets, or even a bit of quick photo editing? Poke tackles it all via plain text messages, in the same chat threads where you talk to friends and family.
This isn’t quite the same as lobbing existential questions at ChatGPT or Claude. If you’re after fast results, or want to put your everyday busywork on autopilot, Poke is the tool to summon. Picture this: ask Poke to ping you when certain emails land in your inbox—maybe from your boss, or a family member. Let it prompt you to grab an umbrella before heading out, or check in about your wellness goals, or simply tell you how the hometown team did last night. It can remind you about daily meds, brief you on news, and streamline even more custom automations you create in plain language—then share with friends if you like.
The company isn’t tinkering at the margins. After raising $15 million last year, the ten-person team secured another $10 million from investors like Spark Capital and General Catalyst, not to mention a roster of influential angels. They’re now carrying a fresh $300 million valuation—a hefty number for such a young product.
Poke emerges at a moment when “agentic AI”—task-oriented digital assistants capable of acting independently—is drawing lots of hype. OpenAI’s recent acquisition of OpenClaw’s creator speaks volumes, as does Nvidia’s CEO preaching the need for a corporate “OpenClaw strategy” alongside their own heavy-duty enterprise agent system. Tech giants are racing, but none quite manage to lower the barrier for everyday users.
For the average person, wrestling with technical installs, dependency errors, cryptic terminal windows, or worrying about security loopholes in complex agent tools is intimidating. OpenClaw’s power brings baggage; few are keen to grant deep system permissions to a program they barely understand.
Poke’s founders sensed this—and wanted to demolish those roadblocks. Marvin von Hagen, co-founder at The Interaction Company (the Palo Alto startup behind Poke), recalls how their previous product—an email-centric AI—showed them something surprising. Even when only designed to help with email, early users began chatting with Poke about everything: medications, the weather, morning jackets, sports scores. Its responsive, almost personable tone made people want to offload more of their lives into its hands.
So the team pivoted. Poke got a broader set of features, smarter behaviors, and much more personality. Getting started with it is laughably easy: drop by Poke.com, hit “Get Started,” hand over your phone number, and you’re off—no downloads, no fuss.
Behind the curtain, Poke decides which AI models to employ based on the task at hand—sometimes leaning on the big providers, other times using open-source options. Von Hagen sees this as a real edge: “Most rivals are part of giant labs, locked to their own models. Meta won’t ever use anything but Meta AI. ChatGPT is bound to OpenAI. We’re model-agnostic by design.”
To function nimbly over chat platforms, Poke relies on Linq—a middleware that lets AI assistants live inside messaging services. SMS and Telegram support is already live. WhatsApp is trickier: last autumn, Meta locked out external general bots. Now, European, Italian, and Brazilian regulators are investigating that move; if Meta changes course (and perhaps drops its steep messaging fees), Poke could find itself back in WhatsApp’s good graces across the EU and South America.

From launch, Poke offers a collection of “recipes”—bundled tools to automate everything from health tracking to budgeting, project planning, travel, school schedules, email triage, and even developer workflows. They slot right into the familiar apps people actually use: Gmail, Google Calendar, Notion, Outlook, Strava, Fitbit, Philips Hue, Sonos, and more. For power users, Poke supports deeper integrations with services like GitHub, Sentry, and Cursor Cloud.
Security is a major concern. The developers built in layered protocols: regular pen testing, tight permission schemes, compartmentalization for both agents and staff. By default, user data stays private; the team only gets a look if you deliberately share specific logs. (Independent audits are still pending, for transparency’s sake.)
Poke’s community has already produced thousands of new “recipes”—the company plans to serve up the best of these in a growing public directory. To spur creators, they’re paying out per-user bounties (roughly 10 cents to a dollar, location depending) for each new sign-up sparked by a recipe.
The pricing structure is flexible, almost conversational. During private beta, users haggled with the AI over their monthly fee—settling between $10 and $30. These days, the price depends on usage; simple tasks may remain free, but real-time jobs (like auto-scanning all incoming emails or live flight statuses) start to accrue cost. The algorithm recommends a price based on complexity and resource intensity.
Profit isn’t the point—at least, not yet. “Monetization is secondary. Our priority is scale, bringing Poke into daily routines for as many users as possible,” says von Hagen. Their growth strategy focuses heavily on creators and influential early adopters demonstrating real-world use cases.
Details about the user base remain vague, though the company says sign-ups have increased tenfold within months. Among the backers are some credible names: Stripe’s Collison brothers, DeepMind’s Logan Kilpatrick, PayPal’s co-founder Ken Howery, Hugging Face’s Thomas Wolf, and a long string of tech entrepreneurs.
Poke hasn’t settled the AI arms race, but it’s brought some genuine human touch—and a welcome ease of use—to a space notorious for technobabble and digital red tape. That alone makes it a launch to keep an eye on.