- The article argues that AI’s primary economic impact lies not in reducing prediction or creation costs, but in lowering “translation costs”—the expense of converting one team’s output into another team’s input.
- Translation costs manifest through meeting times, edits, data reconciliation, and rework; reducing these costs significantly boosts coordination.
- AI achieves this in two ways: extracting structure from unstructured data (emails, PDFs, images, spreadsheets) and using structured data to execute tasks.
- In construction, architects, structural engineers, and contractors use different tools; imposing a universal standard often fails.
- The company Trunk Tools integrates data from Autodesk and Procore to create searchable project records, reducing misalignment. Procore also acquired Datagrid to pursue this direction.
- In the U.S. auto insurance market, CCC Intelligent Solutions dominates due to standardized codes and digitized workflows; high switching costs make it difficult for competitors to challenge them.
- Startup Tractable bypasses the standards war by using AI to read damage photos from smartphones and generate repair estimates compatible with existing systems; by 2023, it processed nearly $7 billion in claims.
- The author proposes three business strategies: becoming an intermediary translation layer (like project44 in logistics), increasing end-to-end accountability (Maersk), or keeping unified internal data and “taxing” access (FedEx).
- In the long run, rapid coordination must be accompanied by governance, accountability, and trust to create a sustainable ecosystem.
📌 AI creates the greatest value by reducing “translation” costs between disjointed systems, enabling coordination without requiring standardized consensus. From construction to insurance, companies like Tractable have processed nearly $7 billion in claims by integrating data instead of changing processes. However, as scale increases, accountability and governance will determine who holds power in this new coordinated ecosystem.
