Workflow
Agentic Search
icon
Search documents
Parallel’s Parag Agrawal: Building a New Web for AI Agents
Sequoia Capital· 2026-08-25 12:00
Industry Trends and Market Dynamics - Parallel Web Systems focuses on scaling agentic search for the agentic web, addressing a market of hundreds of billions of web pages and queries [2][12] - The industry is transitioning from human-driven keyword search to agent-driven search, where agents generate queries that are hundreds to a thousand times more frequent than traditional human searches [62][70] - Web traffic driven by AI is experiencing rapid growth, eventually matching human traffic levels in page reads as background agents continuously monitor the web [71] Investment Opportunities and Potential Risks - Traditional internet monetization models, such as human-centric web advertising and fixed-fee AI lab contracts, face disruption and sustainability challenges as agent traffic replaces human eyeballs [76][80] - Parallel Web Systems implements a differential pricing model utilizing Shapley values to fairly compensate high-value content owners based on incremental value and work quality [81][86][87] - The company established a strategic partnership with Google Cloud to act as a search and grounding provider for Gemini model enterprise agent APIs on GCP [56] Company Financial and Operational Performance - Parallel Web Systems successfully reduced search pipeline compute latency from 3 seconds down to 200 milliseconds while optimizing quality, cost, and latency [28][67] - Search agents utilized by enterprise customers execute between 5 and 20 searches per prompt, scaling up to hundreds of thousands or millions of searches for automated workflows like insurance underwriting and sales data enrichment [21][61][62]