The earned value problem
Every construction project manager faces the same question at every progress meeting: what percentage of the work is actually complete? The answer is usually a combination of site walk observations, subcontractor self-reporting, and educated guessing. It is subjective, inconsistent, and frequently wrong.
The consequences of inaccurate progress reporting are real. Overstating progress leads to over-certification of interim payments — the contractor is paid for work not yet done, which becomes a cash flow problem if the project falters. Understating progress delays payment to subcontractors, damages relationships, and masks genuine achievement. Both erode trust between client, contractor, and design team.
Earned value management (EVM) provides a structured framework for measuring progress: planned value (what should have been done), earned value (what has been done), and actual cost (what it cost). But the "earned value" component — what has actually been built — requires objective assessment of physical completion, and that is exactly what AI can provide.
The traditional approach relies on site walks where the project manager or QS estimates percentage completion for each activity: "Blockwork to first floor — looks about 75% complete." These estimates are subjective. Two people walking the same floor will give different numbers. And on a large project with hundreds of activities, no one can assess everything in a single walk.
