Perplexityが独自の業務特化型AIエージェント「Teammates」を発表

Perplexity's launch of the business‑focused AI agent “Teammates,” announced on November 4 2026, introduces autonomous agents designed to handle complex, ongoing work tasks for enterprises. For website owners, Teammates means a new class of AI bots that will likely crawl sites not just for information retrieval but to perform multi‑step workflows such as data extraction, report generation, and real‑time content updates. Unlike traditional crawlers, these agents maintain state across sessions, so they may repeatedly hit the same pages, increasing traffic patterns that differ from typical search‑engine bots. This raises concerns about server load, data privacy, and the potential for unintended content scraping, especially on sites that host proprietary or regulated information. **Actionable tips:** 1. **Update your llms.txt** to explicitly list “Teammates” as a permitted or disallowed agent, using the exact identifier provided by Perplexity (e.g., `User-Agent: PerplexityTeammates`). 2. **Implement rate‑limiting rules** for persistent agents—configure your web server or WAF to detect repeated requests from the same session token and throttle them to protect resources. 3. **Monitor bot traffic** with llmscentral’s real‑time analytics; set alerts for spikes in requests matching the Teammates signature so you can quickly assess impact and adjust your bot policy as needed.