Rifmont Group is building a private AI system for construction

Rifmont Group, a Vancouver- and New York-based development, infrastructure and advisory group, is developing an internal AI system built on proprietary project data and operating intelligence. The effort signals a bet that construction firms will gain more advantage from private data than from access to generic AI tools.

Why it matters: - Rifmont Group is aiming to turn proprietary project data into a competitive edge in construction. - The approach could help identify procurement issues, permitting delays, cost anomalies and execution risks before capital is committed. - The strategy reflects a broader shift: as AI tools become more common, private datasets may matter more than access to the models themselves.

What happened: - Rifmont Group is developing a private AI environment for construction decision-making. - The system is designed to connect estimating, procurement, permitting, project sequencing, historical costs and execution data. - The company is based in Vancouver and New York. - Founder Jagroop Bhumber said the competitive advantage will come from systems trained on proprietary information unavailable to the broader market.

The details: - Rifmont Group is building the system around construction intelligence gathered from real-world projects. - The goal is an internal intelligence layer, not a commercial AI product. - Rifmont Group has disclosed few details about the system. - The company has not revealed its architecture, training methodology or the scale of the data being assembled. - Bhumber previously outlined plans for Rifmont Group to deploy up to $10 million across real estate, development and infrastructure. - AI is expected to play a role in that strategy, including possible investment in private computing infrastructure, proprietary datasets and specialized technical talent.

Between the lines: - The secrecy appears deliberate, suggesting Rifmont Group sees the data itself as the product. - In construction, execution history and operating intelligence are difficult to copy, even when AI software becomes widely available. - That makes a private system potentially more defensible than a public-facing platform. - Bhumber said the objective is to build technology capable of competing with the market, not for the market.

What's next: - Rifmont Group has not said when the system will launch or how broadly it will be used internally. - More details may emerge as the company expands its broader investment plans in real estate, development and infrastructure. - The key test will be whether the private system can improve decisions enough to justify the investment in data, infrastructure and talent.

The bottom line: - Rifmont Group is betting that in construction AI, the real moat is not the model. It is the private project intelligence behind it.