Field estimation, reimagined

The problem

Selling in the home ran on paper, memory, and mismatched tools. Reps improvised the flow, small mistakes in a complex pricing model cost margin or stalled deals, and the path from inspection to a signed agreement was long enough for momentum to leak away.

The idea

Use AI to carry the busywork of a home visit. Combine transcription, vision, and reasoning so the appointment records and transcribes itself, the conversation fills in the needs assessment, an aerial photo of the house becomes measurements, and a price book turns the scope into a correct, priced quote. The rep runs the conversation; the software builds the paperwork.

What I built

• A recorder that captures and transcribes the appointment as it happens.
• A step that turns the transcript into a filled-in needs assessment, no re-keying.
• Vision over an aerial image of the house to diagram it and auto-calculate the measurements.
• A priced quote built straight from the price book, so the numbers are right by construction.

Impact

• In build now: an AI-assisted appointment that turns the conversation and an aerial photo into a filled assessment and a priced quote.
• Designed to take the busywork off the rep, combining transcription, vision, and reasoning in one flow.
• Measurements read from an image and prices pulled from the book, to remove a whole class of manual mistakes.