Challenges

Riedel Bau [Platinum Partner]

Transforming the way floor slab openings are identified and prepared for construction. The goal is to automatically detect openings for pipes, cables, and technical installations from digital construction plans and use this information to prefabricate formwork elements with concrete 3D printing.
Currently, these openings are often created on-site using manually built wooden boxes, which are later removed and discarded. This process takes time, creates waste, and can lead to errors.
Desired Future State:
Implementation of a system that automatically recognizes and interprets floor slab openings from heterogeneous construction plans and prepares the required data for additive prefabrication.

The challenge: Development of a solution that employs AI to:

• Detect floor slab openings in digital plans.
• Interpret varying labels, legends, scales, and geometries.
• Extract dimensions, position, and slab height.
• Match overlapping plan sections to avoid duplicates.
• Generate structured data for concrete 3D printing.

The solution should be robust enough to handle different plan formats, inconsistent labels, and varying drawing standards from external planning offices.

Challenge Video + Data

will be available shortly before the ConStructAI Hackathon

Riedel Bau [Platinum Partner]

Transforming the way tower cranes are planned and integrated into construction logistics. The goal is to develop an AI-supported workflow that automatically plans and models tower cranes directly within the BIM model.
Tower cranes are essential for high-rise construction, as they move materials between delivery areas, storage zones, and installation points. For efficient crane planning, several factors must be considered, including crane location, height, jib length, lifting capacity depending on reach, and accessibility of all relevant work areas.
Desired Future State:
Implementation of a system that automatically generates technically valid, collision-free tower crane setups in the BIM model.

The challenge: Development of a solution that employs AI to

• Determine suitable crane positions within the BIM model.
• Define crane height, jib length, and lifting capacity requirements
• Check reachability of delivery, storage, and installation areas.
• Detect and avoid collisions with the building, surroundings, and other cranes.
• Generate a technically valid crane layout to support efficient construction logistics.

The solution should reduce planning effort, improve transparency in decision-making, and support faster, safer, and more efficient construction workflows.

Challenge Video + Data

will be available shortly before the ConStructAI Hackathon

Julius Berger International GmbH

AI-assisted checking of building designs for accessibility and workplace regulations. The goal is to analyze 2D plans or BIM models and identify selected compliance issues based on standards such as DIN 18040 and workplace requirements. Today, these checks are often done manually, which is time-consuming, error-prone, and may lead to issues being discovered late in the design process.

Desired Future State:
An intelligent assistant that highlights deviations, risks, and potential improvements directly in the plan or model.

The challenge: 
Development of a solution that employs AI to:

• Analyze 2D plans and/or BIM models.
• Detect selected compliance issues, e.g. corridor widths, door clearances, wheelchair turning spaces, or reachable controls.
• Highlight the affected areas in the plan/model.
• Explain which rule or requirement is affected.

The result should be a working proof of concept that demonstrates how AI can support regulatory design checks in practice.

Challenge Video + Data

will be available shortly before the ConStructAI Hackathon

open bydata

Transforming the way renovation and maintenance works on highways and federal roads are planned. The goal is to identify the most suitable time windows for construction work based on traffic data and, optionally, additional external factors such as weather.
Currently, roadworks often cause major disruptions, including traffic jams, delays, detours for logistics and commuters, increased emissions, and safety risks. A more data-driven approach could help reduce these impacts while still allowing efficient construction progress.
Desired Future State:
Implementation of a system that forecasts suitable time windows for road renovation works and supports logistics-oriented construction planning.

The challenge: Development of a solution that employs AI or data analytics to:

• Analyze traffic patterns on highways and federal roads.
• Identify seasonal, weekly, or daily patterns in traffic volume.
• Forecast low-traffic time windows for planned roadworks.
• Estimate potential impacts on congestion, accident risk, logistics, and commuter traffic.
• Optionally include external factors such as weather conditions.
• Recommend time periods that minimize disruptions while enabling efficient construction work.

The solution should support better planning decisions for roadworks, reduce traffic-related disruptions, and contribute to safer, more efficient, and more sustainable construction logistics.

Challenge Video + Data

will be available shortly before the ConStructAI Hackathon

THWS

Transforming the way informal communication in construction projects is captured, structured, and transferred into project systems. The goal is to use AI to identify relevant information from emails, chat messages, meeting notes, or short text inputs and prepare it for structured follow-up.
Today, important project information is often exchanged informally and manually transferred into schedules, task lists, cost processes, protocols, decision logs, or model-based environments. This is time-consuming, error-prone, and can lead to information being delayed, incomplete, or lost.
Desired Future State:
Implementation of a system that captures informal communication as unverified information, sorts it by topic, supports human verification, and prepares or performs the transfer into suitable project structures.

The challenge: Development of a solution that employs AI to:

• Process unstructured text inputs such as emails, chats, or meeting notes.
• Identify relevant project information across different topics.
• Classify content into categories such as schedule, costs, quality, tasks, responsibilities, decisions, or open points.
• Mark extracted information as unverified until reviewed.
• Provide a simple interface for human verification.
• Assign verified information to suitable target systems or structures.
• Show prototypically how information can be transferred with one click into task management, scheduling, cost, protocol, decision, or model-based environments.

The solution should make visible which information belongs in which system, what still needs review, and what can be passed on as verified project information.

Challenge Video + Data

will be available shortly before the ConStructAI Hackathon