How Mortenson Construction Used AI Progress Tracking to Land Projects Under Budget

The Challenge

Tracking how much of a building is actually built, versus what the schedule says should be built, has always been one of construction's hardest problems. Superintendents historically walked the site, eyeballed progress against drawings, and updated spreadsheets by hand — a process that is slow, subjective, and often days or weeks behind reality. By the time a delay or cost overrun shows up in a report, it has usually already happened.

The Solution

Mortenson Construction integrated an AI computer vision platform (Doxel) that captures the site daily using autonomous rovers and camera-equipped drones, then compares the resulting 3D imagery against the BIM model and project schedule. Deep learning models recognize what's actually been built — framing, MEP rough-in, drywall, finishes — and flag where reality diverges from the plan, without anyone walking the floor with a clipboard.

• Daily 3D site capture replacing weekly or biweekly manual walkthroughs
• Automatic comparison of work-in-place against BIM models and the master schedule
• Real-time variance alerts when a trade falls behind or quality issues appear
• Cost-at-completion and schedule forecasts built from actual measured progress, not self-reported status

The Results

• 38% increase in field productivity, with the project delivered 11% below budget
• Earlier detection of schedule and quality discrepancies, before they became costly
• Reduced delays tied to progress reporting lag
• Improved client confidence through real-time, objective progress reporting

The pattern holds elsewhere: on an 82,000 sq ft healthcare project, Layton Construction used the same approach to cut manual progress-tracking time across six superintendents and project engineers from 60 hours a week to just 3 — freeing up 57 hours a week to redirect toward safety, quality, and trade coordination instead of paperwork.