The ConTechCrew AI in Construction for Business Leaders to Fix Broken Processes & Avoid Deskilling - Jonathan Marsh
Mar 3, 2026
Travis Voss, a construction practitioner focused on building automation, IoT, and workforce development, and Jonathan Marsh, a construction tech consultant specializing in AI, systems integration, and process, talk about AI's role in construction. They explore process mapping, risks of deskilling, IoT and predictive maintenance, systems integration, and rapid prototyping and microcoding trends.
AI Snips
Chapters
Transcript
Episode notes
Context Graphs Can Industrialize Decision Work
- AI can industrialize knowledge work by capturing decision traces across email, Slack, and systems to inform future actions.
- Jonathan Marsh warns automating human decisions risks errors when context or unique one-off choices (like a one-time 20% discount) exist.
Document Processes Before You Automate
- Write down and map processes in detail before automating so AI has accurate examples to scale decisions.
- Jonathan Marsh advises documenting tribal knowledge to inform future decision makers and avoid automating hidden exceptions.
Tribal Knowledge Won't Fully Live In Graphs
- Tribal knowledge lives in workers' heads and often cannot be fully captured by a context graph or logs.
- Travis Voss notes many choices are implicit habits people can't easily explain, so agents should surface exceptions for humans.