Construction project managers were spending 4–6 hours manually reviewing Primavera P6 XER schedule files to identify risks, delays, and critical path issues — a bottleneck that slowed decision-making.
Construction project managers were spending 4–6 hours manually reviewing Primavera P6 XER schedule files to identify risks, delays, and critical path issues — a bottleneck that slowed decision-making.
We built Fortis PMS AI — a React/Node.js web app that accepts XER file uploads and streams AI analysis using the Claude API with Server-Sent Events (SSE) for real-time output.
Critical bugs fixed during build: token history management preventing context overflow, XER metadata routing ensuring correct project data was analysed, and large file chunking to handle schedules with 10,000+ activities.
"What used to take half a day now takes 20 minutes. The AI catches things I would miss."