by Joachim Viktil, Reope
Why do we keep starting over?
Architecture and construction are complex fields. Every project involves countless decisions, from door widths and window placements to floor finishes and rebar layout. Each choice generates valuable data, yet traditionally, we discard most of it, losing insights that could improve future work.
This continual “reinvention of the wheel” is not only wasteful — it also prevents us from learning from experience. But recent advances in AI offer a unique opportunity to change that.
Unlocking AI’s true potential: your own data
General AI (GenAI) tools like ChatGPT, Claude and MidJourney have sparked interest across the AEC industry. While GenAI built on public data can provide benefits, the drawback is that everyone gets access to the same tools trained on the same data. That means the answers you get are likely similar to what your competitors are seeing.
To build long-term competitive advantage, firms must go beyond generic datasets. The real power lies in building AI tools using your firm’s own design data — from past projects, models, and specifications. This makes predictions more relevant and aligned with your style, quality standards, and project history.
By transforming past geometry and parameters into statistics, you can train AI algorithms to predict and now even implement design decisions directly in BIM software like Revit or Rhino. While it sounds complex and expensive, the technology is rapidly becoming more accessible — even for small and medium-sized firms.
The three-step guide to building your AI assistant
As discussed at Digital Construction Week, there are three main steps to building your own AI assistant:
1. Gather and transform your data
Start by identifying valuable, repeated data points — dimensions, materials, configurations — and convert this information from proprietary formats into a usable database for training an AI model.
2. Predict the optimal decisions (the “brain”)
Train AI algorithms using this data to recommend design choices based on what worked well in the past. This could include ideal layouts, materials, or common families used by your team. It can also flag when you’re overusing certain components or missing opportunities for innovation.
3. Implement the decisions automatically (the “hands”)
Using connectors like RevitMCP or RhinoMCP, link your AI model to your BIM software so it can actively generate geometry or optimise model elements. While AI can handle the heavy lifting, human oversight remains essential to maintain quality and intent.
Over time, this process becomes more accurate as feedback loops help your AI assistant continuously improve.
A few potential implications
Agentic AI opens up countless use cases, for example:
-
A project manager quickly summarises design changes in Revit for client updates.
-
A contractor shares a purchase order and receives an optimised installation schedule.
-
An architect receives AI-suggested Revit families for challenging design scenarios.
-
An engineer anticipates software crashes by analysing journal-file patterns.
These examples show what’s already possible when firms combine their data with smart, strategic AI tools.
Balancing speed and human insight
AI is powerful, but not perfect. It operates probabilistically, so while it can speed up workflows, human judgement is still vital — especially for design decisions that impact people’s lives.
As designer Thomas Heatherwick put it:
“Your challenge in designing is finding what the real thing is to solve.”
AI can help us find patterns and possibilities faster, freeing designers to focus on the why behind each choice.
Looking forward: from potential to practice
Digital Construction Week 2025 showcased just how far we’ve come — and how far we still have to go. Integrating AI into your firm’s workflow isn’t a future ambition; it’s a business imperative.
The question now is not whether to adopt AI, but how to do it well — in a way that enhances your competitive edge and creativity.
At Reope, we believe technology should amplify, not replace, human creativity. When you combine your data with the right AI tools, your projects can become smarter, more efficient, and unmistakably yours.