The 1-3% margin stat you mention at the bottom is the key. Technology promises efficiency but at those margins, every implementation cost, every training hour, every software subscription eats directly into survival money. No wonder adoption is shallow.
The submittal rejection rate stuck out to me. 30%+ rejection on 500-2,000 submittals per commercial project is hundreds of revision cycles, each one potentially weeks of delay. The article frames this as a technology problem, but it's really a specification problem upstream. The spec is wrong, incomplete, or ambiguous in the first place — and no project management software fixes that.
What could actually help: AI that reads specifications and flags likely rejection issues before the submittal is even built. Prevention beats management. But that requires integrating with the actual spec documents, not just the workflow software.
The coordination fragmentation (11 different data environments!) is brutal. That's not a technology deficit — that's what happens when every trade buys their own best-in-class tool with no mandate for interoperability. The GC ends up being a human API layer.
Big picture: the industry won't improve until owners stop rewarding low bids that assume perfect execution and start penalizing delays more aggressively. Technology is a band-aid on an incentives problem.
The 1-3% margin stat you mention at the bottom is the key. Technology promises efficiency but at those margins, every implementation cost, every training hour, every software subscription eats directly into survival money. No wonder adoption is shallow.
The submittal rejection rate stuck out to me. 30%+ rejection on 500-2,000 submittals per commercial project is hundreds of revision cycles, each one potentially weeks of delay. The article frames this as a technology problem, but it's really a specification problem upstream. The spec is wrong, incomplete, or ambiguous in the first place — and no project management software fixes that.
What could actually help: AI that reads specifications and flags likely rejection issues before the submittal is even built. Prevention beats management. But that requires integrating with the actual spec documents, not just the workflow software.
The coordination fragmentation (11 different data environments!) is brutal. That's not a technology deficit — that's what happens when every trade buys their own best-in-class tool with no mandate for interoperability. The GC ends up being a human API layer.
Big picture: the industry won't improve until owners stop rewarding low bids that assume perfect execution and start penalizing delays more aggressively. Technology is a band-aid on an incentives problem.