Each dot is one evidence case, coloured by provenance and sized by how close it gets to disclosed money. Hover or tap a dot for the headline; scroll or use the +/− controls to zoom into a cluster. Named project locations are used where the evidence supports them; otherwise the preview identifies a country, company or product-origin marker. Co-located cases are fanned out for legibility, so positions are approximate evidence markers, not a precise GIS layer. Basemap © OpenStreetMap and CARTO.
What do E1–E5 mean?
Geography
Where the evidence sits today, by plotted location.
Provenance
Five Cases have measured results or a published test method. The rest show confirmed use or source-labelled claims.
Sector mix
The built-environment sectors represented so far.
What's in the AI evidence library
The library tracks where artificial intelligence is being used and tested across the built environment. Every entry is a public, source-labelled Case. Together they include live projects, pilots, research tests and relevant industry comparisons. Roughly half of the sources are the suppliers themselves, which is exactly why every Case shows who is making the claim.
It currently covers 85 public-map Cases across the UK, United States, Europe and Asia Pacific, spanning construction, water, transport and rail, energy, real estate and planning.
- Construction & AEC: AI for project risk, rework, productivity, robotics and capital-project controls.
- Water & flood: network monitoring, leakage, wastewater and flood resilience.
- Transport & rail: inspection, scheduling and infrastructure operations.
- Energy, buildings & planning: asset management, smart buildings and planning automation.
How is AI being used in the built environment?
Across construction, water, transport, energy, real estate and planning, for project-risk and rework detection, asset inspection, network monitoring, scheduling, planning automation and more. Each case in the library is a real, source-labelled example rather than a generic prediction.
What does each evidence case include?
The company or project, sector, geography and lifecycle stage, plus a 38thMove provenance label (Measured, Confirmed or Claimed) and an E1–E5 score for how close it gets to disclosed money. The underlying source trail and analysis inform 38thMove advisory work.
How does the library support the advice?
It helps test whether a claim is isolated or recurring, who is making it, how strong the public evidence is and what a leader should ask before acting on it.
Evidence into decisions
A map of Cases is useful. It is not a recommendation.
The map shows all 85 published Cases with their provenance and value read. Advisory work goes further: connecting the evidence to the decision, organisation and consequences in front of you.
- Context: what experience says about the market, organisation and practical constraints.
- Challenge: what the source trail supports, where the claims are weak and what remains unknown.
- Action: what to proceed with, test, narrow, pause or stop.
Discuss a decision →
Senior, independent advice grounded in experience and disciplined public evidence.