AI design & operations platform for contractors

The situation

Outdoor construction loses deals in the gap between the site visit and the proposal. A contractor measures a backyard, goes away, draws something over the next fortnight, sends a quote — and by then the homeowner has cooled off or taken three other bids. The brief was to collapse that gap to a single visit.

What we built

• An AI design pipeline that takes a photo of a real site and returns a photorealistic render of the proposed work, leaving the existing house untouched and changing only the outdoor space
• Product selection drawn from real manufacturer catalogues, so what the homeowner approves is something that can actually be ordered
• Style transfer and inpainting for targeted changes, plus walkthrough video generation
• A design-to-quote pipeline converting the approved render into a line-item estimate with material quantities and smart pricing
• Proposals, PDF contracts and e-signature capture, so a deal closes on site
• Payments and invoicing through Stripe, with financing and automated payment reminders
• Project management, workforce scheduling and financial reporting in the same workspace

Engineering notes

The hard constraint was credibility. A render that beautifies the whole house is useless for selling a patio — the homeowner stops believing it. The pipeline had to change only the outdoor area while leaving roof, siding, windows and walls exactly as photographed, and it had to place products the contractor can really buy rather than generic shapes. That requirement, not the image generation itself, is what shaped the architecture: the design and the quote read from the same product selection, so the price can never describe something different from the picture.