What I've Learned Building AI for Construction

Bassem Hamdy · CEO, briq · Mar 2026

Personal reflections from briq's CEO on the challenges, surprises, and lessons of building AI specifically for the construction industry.

When I started briq, I thought the hard part would be the AI. I was wrong. The AI — the models, the algorithms, the infrastructure — is genuinely challenging, but it's a known quantity. The truly hard part has been understanding construction deeply enough to build AI that actually works in the real world. That journey has taught me lessons I want to share, because I think they're relevant for anyone building or buying technology for this industry.

The first lesson is that construction is far more complex than outsiders appreciate. I came from a technology background, and I assumed construction was a relatively straightforward industry that just hadn't adopted good technology. I was humbled quickly. The interplay between contracts, regulations, labor agreements, site conditions, material markets, and project-specific requirements creates a combinatorial complexity that rivals any industry I've encountered. Respecting that complexity is the prerequisite for building anything useful.

The second lesson is that data in construction is messy — and that's not going to change. Early on, we spent a lot of effort trying to get customers to standardize their data before using our platform. That was the wrong approach. Construction data will always be messy: handwritten notes, inconsistent formats, varying naming conventions, incomplete records. The AI has to work with the data as it is, not as we wish it would be. This realization fundamentally shaped our engineering approach.

The third lesson is that trust is earned in inches. Construction professionals are rightly skeptical of technology claims. They've been burned by vendors who promised transformation and delivered a half-built portal. When we deploy AI that processes invoices, the customer doesn't just take our word that it works — they verify every output for weeks before trusting the system. We design for this trust-building process explicitly, providing full transparency into how the AI reached every decision.

The fourth lesson is that integration is not a feature — it's the product. I used to think of integration as a secondary concern, something you handle after building the core AI capabilities. I now believe integration is the core capability. If your AI can't connect to the customer's ERP, project management system, and document repositories, it doesn't matter how good the AI is. It can't access the data it needs and can't deliver results where they're needed. We've invested more in our integration platform, GLYPH, than in any other part of our technology stack.

The fifth lesson is about the importance of domain specificity. We tried using general-purpose models early on and found that they failed on construction-specific tasks in ways that were subtle and dangerous. A general model might correctly extract an amount from an invoice but misinterpret the retainage calculation. It might read a change order but not understand the contractual implications. Domain-specific training isn't optional — it's the difference between a system that works in demos and one that works in production.

The sixth lesson — and perhaps the most important — is that the construction industry is ready for this transformation. The skepticism that I encountered five years ago has largely given way to urgency. The labor shortage, margin pressure, and competitive dynamics have created a pull for AI solutions that didn't exist when we started. Construction companies don't need to be convinced that they need AI — they need to be shown AI that actually works for their specific needs.

Building AI for construction has been the most challenging and rewarding work of my career. Every day, I learn something new about this industry, and every day, I'm more convinced that AI will transform how construction companies operate. We're still early in this journey, but the progress we've made gives me enormous confidence in what's ahead.