OpenAI Unveils Always-On AI Agent 'Dots'

OpenAI's unveiling of always‑on AI agents called “Dots,” which can proactively use a computer and connected apps to research, draft documents, write software and handle other tasks on a user’s behalf, signals a new wave of autonomous web interactions. For website owners, Dots means AI traffic that is no longer limited to single‑prompt queries; these agents will continuously browse, fill forms, scrape data and even trigger downstream API calls. Because Dots operate as persistent sessions, they can generate sustained load spikes, bypass traditional bot detection that looks for rapid request bursts, and potentially interact with paywalls or personalization engines in ways that mimic genuine user behavior. Ignoring this shift could lead to skewed analytics, inflated engagement metrics, and unintended exposure of premium content. **Actionable tips:** 1. Update your llms.txt to explicitly list “Dots” as a disallowed agent or require a custom header (e.g., `X-Dots-Client: true`) so you can filter or throttle its requests. 2. Deploy a real‑time bot‑tracking solution that flags long‑duration sessions with repetitive cross‑site navigation, then route those sessions to a honeypot endpoint for deeper inspection. 3. Implement rate‑limiting rules based on session length and concurrent connections rather than just request frequency, ensuring that always‑on agents cannot monopolize resources.