AI for AEC — 5 Insights for 2025

https://unsplash.com/photos/photo-of-high-rise-buildings-fglv83gZaxU?utm_content=creditShareLink&utm_medium=referral&utm_source=unsplash

Bringing together the worlds of architecture and software is a dance between clarity and complexity.

This year, that dance took some unexpected turns. Professionally, my company, DBF, had to evolve quickly to meet disruptive changes in technology and market demands. And on a personal note, I returned to North America after a decade in Asia and jumped back into the world of design education as a part-time lecturer at the University of Waterloo at a time of profound challenges and change.

Looking ahead to 2025, I want to share my 5 reflections from the past year

In Short, They Are:

1. GenAI as UI
2. “Design Once, Build Many” vs. “Localization”
3. User trust vs AI
4. AI as mindset vs AI as toolkit
5. Time ≠money

In Long Form, They Are:

1-GenAI as UI

Generative AI shows great potential for lowering the barriers to entry and revitalizing legacy software.

Anyone who has attempted to learn 3D modeling or drafting tools knows the frustration: outdated interfaces, steep learning curves, and essential features hidden behind archaic menus.

However, working with clients like Takenaka Corporation this year, I witnessed firsthand how generative AI can act as a powerful “translator” for these legacy systems. While the results aren’t perfect, the potential is undeniable.

Key Takeaway:

• Rather than replacing trusted tools, we can enhance their value through AI-driven interfaces that improve usability and extend their lifespan.

2- “Design Once, Build Many” vs “Localization”

This year, “localization” became the defining buzzword in many discussion I’ve had with consultants. As my business partner likes to say, The building industry faces global challenges but requires local solutions.

Compliance codes, business requirements, and climate considerations vary so widely that one-size-fits-all software often falls short.

Using GenAI, we are currently working on new systems that ingest regional regulations and integrate seamlessly with our existing generative building design tools. These next-generation tools could go beyond simple compliance checks — they suggest and even modify designs to meet new or emerging requirements.

Key Takeaways

• While “design once, build many” is the goal of many projects, especially facilities, generative AI could be the breakthrough that enables truly localized computational design tools, identifying the unique challenges and requirements of each site and client.

3 —AI vs User Trust

While most firms I talk to excited are about AI’s possibilities, many remain cautious are still rightly cautious.

I belive transparency is often the biggest barrier; Despite all its potential, AI requires careful validation and cross checking;Without clear explanations of how AI makes decisions or safeguards data, adoption will be impossible.

I see this as as much a design and people challenge as a technology one. We need to design products in a way that make AI processes visible and verifiable. Features that show how decisions are made and enable cross-checking have been instrumental in building confidence.

Key Takeaway

• Building trust in AI is a product design challenge as it is a technical one.
• Transparent systems that empower users to validate results are essential for widespread adoption.

4 —AI as mindset vs AI as toolkit

Stepping back into the unversity this year made one thing clear: education is still coming to grips with how to respond the fresh wave of AI. Most architecture schools focus on teaching tools but fail to address the critical, ethical, and creative thinking skills that underpin their use holistically

At the University of Waterloo, I’ve approached teaching through “learning by doing” to perpare students to work with AI; students need hands-on experience with AI in simulated projects, learning to question its results and iterate with it as a collaborator.

Key Takeaway

• To prepare future designers, we must teach them how to think critically about AI and its applications — blending technical skills with reflective design.

5 —Time ≠money

In 2024, I spoke with many real estate developers and government agencies who see AI and enterprise software as fundamentally a cost-saving measure. In turn, this view pressures on constulants and firms to innovate to stay on top. However, one of the biggest questions this year has also been: How can we monetize AI or other emerging tech, without undermining the traditional service models that AEC firms and consultants rely on?

Key Takeaway

• By treating AI as a growth engine rather than a competitor, firms can unlock new revenue streams and remain competitive in an evolving market.
• After all these years, I remain the bullish on the opportunity for AI and other emerging tech to disrupt the industry to the stagnancy of traditional AEC business models.

Love to hear your thoughts?

What AI trends do you see shaping your work in the builidng / urban planing sector? What lessons have you learned from adopting or developing new tools?

About the author

Sayjel Patel is the CTO and Co-Founder of Digital Blue Foam (DBF), dedicated to transforming the Architecture, Engineering, and Construction (AEC) industry for the digital era. An MIT-trained architect and technologist, Sayjel merges over a decade of experience as an entrepreneur, designer, software developer, researcher, and educator, having taught in Singapore, the UAE, and Canada.