Discovery Friction
The effort required for AI systems to locate, parse, reconcile, infer, and validate information before including entities in consideration sets.
Description
Discovery friction consists of five components: intent resolution friction, retrieval friction, comparison friction, representational friction, and verification friction. High friction creates computational cost and may exclude entities from consideration. Structured representation reduces friction.
Related Concepts
Related Research
Discovery Cost Collapse
The legacy web was built on friction: navigation costs, comparison costs, advertising competition, duplicated inventory, and retrieval inefficiency. These inefficiencies were not bugs—they were features that created economic opportunities for intermediaries, search engines, and aggregators. This paper argues that AI-mediated markets may fundamentally compress discovery friction through structured representation, machine-readable interoperability, and reasoning-based matching. As AI systems increasingly mediate discovery, comparison, and recommendation, the economic center of the web may shift from attention acquisition toward representational efficiency and reasoning quality. We introduce a formal framework for discovery friction, define the transition from retrieval economies to understanding economies, and analyze the structural economic implications of AI-mediated discovery compression.
The Web Retrieval Cost Report 2026
The Web Retrieval Cost Report 2026 measures the effort required for AI systems to locate, parse, reconcile, infer, and validate information from web sources before producing answers to property discovery queries. When property information exists only in fragmented web pages, listings, PDFs, and portal content, AI systems must perform additional work before they can compare or select properties. This report establishes that structured property records reduce retrieval cost by making relevant attributes directly accessible, connecting web search efficiency to representation quality. Through observation of AI-mediated property discovery across 50 markets, thousands of AI responses, and systematic evaluation of retrieval sessions, we demonstrate that retrieval cost is a measurable component of AI discovery efficiency.
The Property Retrieval Failure Report 2026
The Property Retrieval Failure Report 2026 measures and explains a phenomenon increasingly observed in AI-mediated property discovery: a property may exist online and still fail retrieval. This report establishes Retrieval Failure as a measurable phenomenon, distinguishing between Information Availability, Information Retrievability, and Information Usability. Across 50 markets, 12,000 observed AI responses, and 8,000 evaluated retrieval sessions, we document how properties fail AI-mediated selection because required attributes are unavailable, fragmented, ambiguous, inconsistent, or not represented in machine-readable form.