ConcreteOS
An AI design assistant and quoting platform for architectural precast concrete. Builders describe a facade, wall, or paver in conversation; the AI recommends products and finishes, estimates manufacturability and carbon, and drafts the quote — with a full operations dashboard for admins.
An architect or builder snaps a sketch, describes a facade, or uploads a BIM file; the AI grasps the requirement, picks the right precast family and finish, estimates cost, manufacturability, lead time, and carbon footprint, then drafts a preliminary quote. Admins review, adjust the assumptions, and approve.
What it solves
- Specifying and quoting concrete by hand is slow; designers have to call the factory to confirm options.
- Estimating manufacturability, lead time, and carbon takes specialized factory know-how.
- Quoting is manual and error-prone; admins must pull the scope out of conversations and sketches by hand.
Impact
Describe or upload sketches, BIM, PDFs
Product family, finish, manufacturability
Cost, carbon, lead time, confidence scores
Architecture
Data flow
- Designer describes requirement (text or upload sketch/BIM)
- AI extracts brief (dimensions, finishes, constraints)
- Clarifying questions → structured requirement
- Product recommendation (family + finish)
- Estimate: cost, lead time, manufacturability, carbon
Confidence scores included
- Draft quote → workspace (conversation + files)
- Admin review → adjust assumptions → approve
- Binding offer
Engineering decisions
Conversational brief extraction over rigid forms
Designers describe the requirement naturally; the AI asks clarifying questions (height, finish, airflow, color) and builds a structured brief incrementally.
Rule-based estimator + AI for confidence
Deterministic rules (product specs, finishes, capacity) provide reliable cost & lead time; AI adds context awareness and confidence scores so admins know when to override.
Dual-mode: mock + production
Mock mode (default) runs the full UX without Firebase or AI credentials — perfect for demos and testing. Flip env vars to enable real Firebase, Cloud Functions, and LLM calls.
Admin approval workflow with assumption editing
Every preliminary quote is "subject to internal review." Admins review, edit cost assumptions, adjust confidence scores, and approve — before the quote leaves the factory.
Gallery
