This is a proof-of-concept version of the product developed for Peikko company during Junction2024 Hackathon.
Peikko sells industrial connections for construction, such as beam-to-column connectors, and faces a common problem:
- Customers provide IFC files containing complex building information.
- Choosing the right Peikko products requires expert analysis.
- Sales engineers invest time in recommendations, often without immediate sales.
The goal? Automate and enhance the recommendation process with an AI-driven tool that improves efficiency for both customers and Peikko’s sales team.
We built a Blender-based tool that:
✅ Visualizes the IFC file as a 3D model.
✅ Analyzes the structure and identifies places where Peikko connections are needed.
✅ Recommends the best products based on the structure and Peikko’s catalog.
✅ Provides an AI assistant to answer customer queries with real-time contextual knowledge.
- The IFC file is imported into Blender to reconstruct the building model.
- The system analyzes structural elements to detect areas where Peikko connections are necessary.
- Key points are highlighted for the customer to explore further.
- The user clicks on a highlighted area, and the system suggests suitable Peikko products.
- The recommendation is based on building specifics, Peikko’s product catalog, and industrial manuals.
- The customer can interact with a smart AI assistant to:
- Ask about specific product details.
- Understand why a recommendation was made.
- Get installation or compliance guidance.
- The chatbot uses MeiliSearch based knowledge base to provide context-aware responses.
This project was developed as part of Junction 2024 Hackathon.
Team Members:
For inquiries, feel free to reach out or check Peikko’s official resources.