Adjusters photograph the damage into a dedicated WhatsApp channel from the site. GPT-4 Vision reads the photos, extracts damage details, drafts the inspection report in the firm's template, and files it against the correct claim — ready for review the same day.
Typical results
Same-day
Report delivery
+30%
Inspection capacity
20 min
Adjuster review time
100%
Photos filed to claim
Figures reflect typical results for this kind of automation.
The challenge
Adjusters inspected three or four sites a day but wrote reports at night or on weekends. Photos sat on phones, details faded from memory, and finished reports reached insurers five days after inspection on average. Slow reports meant slow settlements, frustrated insurers, and adjusters burning out on paperwork instead of fieldwork.
Before and after
Report turnaround
−80%Writing time per report
−83%Inspections per week
+32%The automation
We set up a WhatsApp channel where adjusters drop site photos and a short voice note tagged with the claim number. A Make scenario passes everything to GPT-4 Vision, which identifies the damaged property, damage type, severity and visible cause, then drafts a full inspection report in the firm's own template. The draft files itself against the right claim in the case system, photos attached and captioned, awaiting the adjuster's review and sign-off.
How it works
- 01
Snap on site
The adjuster photographs the damage and records a short voice note in the WhatsApp channel, tagged with the claim number.
- 02
AI reads the scene
GPT-4 Vision identifies the property, damage type, severity and visible cause from the photos.
- 03
Draft in the template
The report is drafted in the firm's inspection template with captioned photos embedded.
- 04
Filed to the claim
The draft attaches itself to the correct claim in the case system automatically.
- 05
Review & sign
The adjuster polishes and signs the report — same day, from anywhere.
Starts from
- Site photos
- WhatsApp channel
Delivers
- Drafted report
- Filed to claim
- Adjuster sign-off
The impact
Reports that took five days now reach insurers the same day — adjusters review and sign a ready draft in about 20 minutes instead of writing for two hours. The firm handles 30% more inspections with the same team, and evening paperwork has all but disappeared.
Questions about this automation
How does the AI turn site photos into an inspection report?
The adjuster posts photos and a short voice note to a WhatsApp channel, tagged with the claim number. Make sends them to a vision-capable AI model that identifies the property, damage type, severity and visible cause, then fills the firm's own inspection template with the findings and captioned photos.
Does the adjuster still review the report before it goes to the insurer?
Yes. The AI only produces a draft. It is attached to the correct claim in the case system and marked as awaiting review, and nothing reaches the insurer until the adjuster edits and signs it. Photos the model cannot read clearly are flagged in the draft so the adjuster can add the missing details.
What is needed to set this up for a loss adjusting firm?
A WhatsApp number for the field channel, the firm's current report template, and access to the case or claims system where reports are filed. It runs as one path from photos to a filed draft, and simple automations like this are usually live within days, often tested with a few adjusters first.