Poke raised $10M more at a $300M valuation, letting consumers run AI agents via text message without downloading any app.
Poke, a 10-person AI startup backed by Spark Capital and General Catalyst, publicly launched an AI agent accessible via iMessage, SMS, Telegram, and WhatsApp. The agent handles calendar management, health tracking, smart home control, photo editing, and custom plain-text automations. The company closed a new $10M round on top of a $15M seed, reaching a $300M post-money valuation. Notable angels include founders from Stripe, Cognition, Vercel, PayPal, Dropbox, and Hugging Face.
Poke's real technical bet is that the messaging layer — not a custom app — is the lowest-friction interface for agentic tasks. For developers building AI tools, this validates SMS/messaging webhooks as a serious deployment surface. The architecture implies persistent context, integration with third-party APIs (calendar, health, smart home), and plain-text automation parsing — all without a mobile app install.
Prototype a single-purpose SMS AI agent this week using Twilio and an OpenAI function-calling endpoint — pick one high-frequency task (e.g., daily standup summary) and measure completion rate vs. your current app-based flow.
Suggested exercise · not independently tested. Access requirements and costs may vary.
Open a Python environment with the Twilio and OpenAI SDKs installed
Paste this starter: `from twilio.rest import Client; from openai import OpenAI; client = OpenAI(); response = client.chat.completions.create(model='gpt-4o', messages=[{'role':'user','content':'Summarize my top 3 priorities today and reply in under 50 words'}]); print(response.choices[0].message.content)` — swap in a real system prompt tied to your use case
Wire the output to a Twilio SMS send call and text yourself the result
Tags