San Francisco‑based The Interaction Company introduced Poke, a text‑based AI assistant that lets consumers automate everyday tasks without downloading an app. The service went live to the public in March, offering a personal assistant that operates through familiar messaging platforms such as iMessage, SMS and Telegram, with limited WhatsApp support in markets where the channel is available.

Unlike general‑purpose chatbots that answer questions, Poke is built for action. Users can ask the agent to flag specific emails, remind them to bring an umbrella, track fitness goals, or even fetch sports scores. The system relies on plain‑text “recipes” – pre‑made or user‑generated automations – that can be installed with a single click and then run automatically in the background.

Behind the product, Poke runs on the AI model best suited to each task, pulling from major providers or open‑source alternatives. The startup leverages Linq to embed the assistant within messaging apps, allowing it to interact with services like Gmail, Google Calendar, Outlook, Notion, Strava, Fitbit and Philips Hue. Developers can also tap into integrations with tools such as PostHog, Webflow, Supabase, Vercel, GitHub and Sentry.

The company’s security model is layered, featuring regular penetration testing, permission‑limiting for both agents and staff, and token encryption that only the user can unlock. According to co‑founder Marvin von Hagen, the team cannot view token contents unless a user explicitly shares logs or analytics.

Poke’s funding round adds $10 million to the $15 million seed round raised last year, bringing its post‑money valuation to $300 million. Investors include Spark Capital, General Catalyst and a slate of angels such as Stripe founders John and Patrick Collison, PayPal co‑founder Ken Howery, Dropbox co‑founder Arash Ferdowsi, and Hugging Face co‑founder Thomas Wolf.

Pricing is usage‑based. Basic automations that do not require real‑time data remain free, while real‑time inference – for example, monitoring incoming emails or checking flight statuses – incurs a charge determined by the system’s cost model. During beta, users negotiated monthly fees ranging from $10 to $30; the company now tailors pricing to individual usage patterns.

Since launch, Poke’s user base has reportedly grown tenfold, though exact numbers were not disclosed. The startup encourages creators to share recipes, offering modest payments of ten cents to a dollar per new sign‑up generated by a shared automation. The goal, von Hagen says, is to embed Poke into everyday life rather than focus on immediate profitability.

Industry observers note that Poke arrives as demand for agentic AI surges, with competitors like OpenAI’s OpenClaw and Nvidia’s enterprise‑grade offering targeting similar markets. Poke differentiates itself by eliminating the need for software installation, terminal commands or dependency management, lowering the barrier for non‑technical users who might otherwise shy away from complex AI agents.

With its multi‑platform approach, flexible pricing and a growing library of recipes, Poke aims to become the go‑to personal superintelligence for a billion users, positioning itself as a practical alternative to more heavyweight AI solutions.

Used: News Factory APP - news discovery and automation - ChatGPT for Business