Overview
- Cloudflare and OpenAI have started an early-stage research pilot to share real-time network signals so OpenAI can test whether those signals improve how AI systems find and index open-web content.
- The pilot will feed signals such as page-change events, content freshness indicators and traffic-quality measures into OpenAI’s search and answer stack to evaluate effects on accuracy and timeliness of responses.
- Cloudflare says its global network fronts more than 20% of the web, which it argues gives the company a unique, large-scale view of when pages change and how traffic behaves in real time.
- OpenAI will supply frontier models, a large-scale search and answer system, and real user queries to measure whether signal-driven crawling reduces wasted crawling and surfaces newer, relevant pages.
- The test sits beside Cloudflare’s recent policy and product moves on crawler defaults and Pay Per Use pilots and raises questions about who will control attribution, permissions and payments for web content.