An existing building whose materials no one knows exactly anymore. A construction site where no one knows for sure which materials have already arrived today. An architectural plan that must be translated by hand into a fabrication specification.
In many areas, the construction industry still relies on processes that are time-consuming and frustrating—and this is exactly where our second AI Challenge in Karlsruhe came in. Together with Fraunhofer IOSB, the City of Karlsruhe’s Energy Forum, TALKCONCRETE, Objektkultur, and the CyberForum, we brought together AI and construction experts to develop concrete project ideas based on four thematic areas.
What a Building Really Has to Offer
It is currently difficult to reliably estimate the actual benefits of reusing building materials—existing buildings are often poorly documented, and new buildings are rarely planned with a time horizon that matches their actual service life. This is where the “Ö=C=Ö” project concept comes in: An AI agent builds a materials database from available data sources, estimates missing information in product data sheets, and uses this to generate concrete, data-driven recommendations for architects—solutions that are both economically and environmentally sound, rather than a compromise made after the fact.
When no one knows what has already arrived at the construction site
Construction companies with multiple active job sites at the same time are familiar with the problem: Suppliers deliver materials, but whether and where they have actually arrived often remains unclear if no one is directly on site. The “DeliveryScan AI” project uses cameras—supplemented by RFID or color coding—to record deliveries and automatically reconciles them with orders and bills of materials. A related problem concerns the condition of the construction site itself: “SiteSense AI” guides employees through the on-site documentation using voice prompts, compares photos and details with target data, and thereby reduces the number of on-site visits required by construction management.
When every technical model has to be rebuilt
A planning model here, a cost estimation model there, an as-built model based on point clouds—in practice, the data for a construction project is often stored in multiple systems that don’t communicate with one another. As a result, domain-specific models are frequently created twice, which slows down the bidding process. The “AI Mapper” project concept creates a common target model onto which existing domain-specific models are mapped—either automatically or with AI support. In clear-cut cases, this process runs deterministically; in cases of ambiguity, the AI suggests mappings that subject matter experts review and approve—control remains with humans, while speed comes from automation.
When an architectural plan needs to be turned into a manufacturing specification
Before a prefabricated component can be produced, an architectural plan must be translated into a specification suitable for manufacturing—a task that, until now, has been performed manually by technicians. The “AI Elements” project combines computer vision for plan analysis with a generative AI model that automatically generates an optimized cutting plan from the architectural plan—taking into account factors such as material efficiency or CO₂ emissions.
What Matters in the End
Five project ideas, four thematic areas, one common thread: AI never makes the final decision, but rather does the groundwork—it collects data, compares it, and makes suggestions, while architects, engineers, and construction managers retain control. Whether Ö=C=Ö, DeliveryScan AI, SiteSense AI, KI-Mapper, and AI Elements will become concrete projects is now being decided within companies in the construction industry.
Learn more about the AI Challenge “Creative Workshop for the Construction Industry”
Here's what happens next
On December 8, representatives from the construction industry will meet with research institutions and providers of AI solutions at the Center for Virtual Engineering (ZVE) in Stuttgart.
Together with the AI Alliance Baden-Württemberg, Fraunhofer IOSB, Fraunhofer IRB, and TALKCONCRETE, we are organizing a day full of inspiration, workshops, and networking opportunities for companies, startups, associations, and research institutions that want to take concrete steps to leverage the opportunities offered by artificial intelligence in the construction industry.
Creative Workshop for the Construction Industry
In our workshop “Creative Workshop for the Construction Industry,” AI and construction experts, along with interested participants, collaborate to develop innovative ideas and solutions for making construction safe, efficient, and less burdensome—all while incorporating smart technology. The results are documented in the form of project profiles based on the PAISE® methodology developed by Fraunhofer IOSB, the key content of which is subsequently published as project summaries.
Organizers:
AI Alliance Baden-Württemberg | Fraunhofer IOSB | Fraunhofer IRB | TALKCONCRETE | Stuttgart Economic Development Agency