Email & photo intake automation

The problem

Property details arrived across emails, attachments and photos. Preparing a record and report meant re-entering the same information.

What I built

I built Scout to extract details, analyze photos, generate repair estimates and PDF briefs, and create CRM records.

The result

The workflow ran end to end, from inbox to estimate, brief and CRM record. Well over the 17 photos in the documented test went through it in practice and the estimates were usable for offer decisions; I never measured accuracy as a percentage.

How it fit together

Inbound emails and attachments were parsed into a structured record; photos ran through analysis that fed a repair estimate; the record, estimate and evidence were rendered into a PDF brief and written to the CRM, with estimates marked for operator review.

Structured schemas connected text and photo inputs to report generation and CRM delivery. Estimates remained available for human review.

Decisions and tradeoffs

Show the evidence behind each estimate, not just the number.

A longer brief, but an operator can disagree with a specific photo call instead of the whole estimate.

One workflow from inbox to CRM rather than separate tools.

More coupling in one codebase; no re-entry of the same details three times.

Photo analysis supported estimates for an operator to review. The evidence behind an estimate mattered as much as the number.