Photo capture + AI scope
Snap the room. V104 reads the conditions and suggests scope.
This is the second wedge that pulls V104 ahead of generic estimating tools. You walk into a job, hold up the phone, take a few photos, and the AI suggests scope based on what it sees. Combine it with voice and you've got a usable estimate before you leave the driveway.
When to use it
How the AI analyzes the photo
V104 looks at the image and identifies fixtures, finishes, surface types, visible damage, and likely scope. It doesn't measure (yet) — but it'll flag things like "tile floor, appears cracked", "older vanity, single sink", "popcorn ceiling", or "drywall water staining near corner."
From those observations it generates suggested line items: Demo existing vanity. Replace tile flooring. Skim and re-finish ceiling. Drywall repair, 4 sf.
Reviewing AI-suggested scope
Every photo-derived line item shows up flagged with a small camera icon, so you know it came from AI vision rather than your own input. Review them carefully before pushing into the estimate:
- Confirm the AI got the surface area right (it estimates, but doesn't measure)
- Check fixture counts — easy to miss a second sink in a wide-angle shot
- Add anything the camera missed (subfloor condition, electrical, anything outside frame)
- Reject items that don't match the actual scope of work
Photos auto-bind to the estimate ID
Every photo you take in V104 is automatically tagged with the estimate ID, the lead, and the timestamp. They live in the estimate's photo gallery and travel with the job into Job Cockpit. If a client ever pushes back on scope or pricing, the photos are right there in the record.
Best practices
- Good light. Daylight beats flash for AI accuracy.
- One subject per shot. Wide angles confuse the model.
- Multiple angles per room. The AI gets sharper with more views.
- Tag photos with a short voice note for context the camera can't capture.