Research & Benchmarks

How we build open data infrastructure for the agent economy. Local LLMs, entity resolution, and classification at scale.

Featured
Agent Readiness: The Attribute Nobody Has
We scanned 14,000+ business websites for AI agent readiness. Six categories, 10 signals, 0-100 score. Average score: 43 out of 100. Only 3.7% grade A. See who's ready and why it matters.
Cam Fortin · April 2026
Benchmark
The Prompt That Changed Everything: B2B Classification from 17% to 77%
We tested 5 local LLMs with 3 prompt strategies on 30 real Boise businesses. Chain-of-thought reasoning took accuracy from 17% to 77%. gemma4:e2b hits 73% on name alone.
Cam Fortin · 2026
Originally published on Product Hacker
Benchmark
Small Models, Big Questions: The Real Bottleneck Isn't the Model
5 models, 5 rounds, 30 real businesses. We went from 0% NAICS accuracy to 83% by splitting semantics from codes. The taxonomy was the bottleneck, not the model.
Cam Fortin · 2026
Originally published on Product Hacker