For Insurance Claims Adjusters ·
What you'll accomplish
By the end of this guide, you'll be using Claude Pro or ChatGPT Plus to analyze damage photos, generate organized photo captions, identify damage items for your claim file, and flag things you may have missed on your inspection. This turns your post-inspection photo review from a 60-minute task into a 5-minute one.
What you'll need
Both Claude Pro and ChatGPT Plus support image analysis. Claude is generally better at detailed descriptions and understanding context across multiple images; ChatGPT handles large photo batches well.
For this guide, we'll use Claude Pro, but the steps are nearly identical in ChatGPT Plus.
Go to claude.ai, log in, and click New conversation. Set context first:
I'm an insurance claims adjuster who just completed a property inspection. I'll share damage photos with you. Please analyze each image and help me document the damage for my claim file.
Click the paperclip icon in the chat input. Select 3-5 photos at once (or one at a time — both work).
Wait for the photos to upload. You'll see thumbnail previews appear above the text input.
What you should see: Small photo thumbnails in the chat input, confirmed by Claude saying something like "I can see [X] images attached."
Troubleshooting: If photos are too large to upload (over 5MB each), compress them first using a free tool like tinypng.com or by using your phone's "Share → Reduce File Size" option before sending.
Type your prompt and press Enter. Here are the most useful prompts:
For comprehensive damage documentation:
Please analyze these property damage photos and provide: 1) A description of the damage shown in each photo, 2) The likely cause of damage if identifiable, 3) What trade work would be needed to repair (e.g., drywall, roofing, flooring), 4) Any damage that appears severe or structural, 5) Anything you notice that I should investigate further.
For quick captioning:
Write a brief, professional photo caption for each image that describes the damage visible. Format: "Photo [number]: [damage description]." Use insurance/construction terminology where appropriate.
What you should see: A detailed response analyzing each photo, organized by photo number, with specific damage descriptions and observations.
Copy the AI-generated captions and analysis. Paste them into your claim documentation system as the basis for your photo descriptions. Edit any descriptions that are inaccurate based on your firsthand knowledge from the inspection.
For Xactimate estimates: review the AI's list of identified damage items against your scope to make sure you haven't missed any line items.
For property damage documentation:
Analyze these property damage photos. For each: describe the damage, identify the type (fire/water/impact/etc.), note severity (cosmetic/moderate/severe), and list repair trades needed.
For roof damage:
Review these roof damage photos. Identify: type of damage (hail/wind/impact/age), affected areas and coverage percentage, granule loss indicators, and whether structural damage appears present.
For auto damage:
Analyze these vehicle damage photos. Describe damage to each area (front/rear/door/etc.), estimate damage severity (minor/moderate/major), and note any structural or safety concerns visible.
For water damage:
These are water damage photos from a property claim. Identify: source of water if visible, scope of affected areas, any secondary damage (mold, structural), and whether the damage pattern is consistent with a sudden loss versus gradual seepage.
For checking your scope completeness:
I've completed my Xactimate scope. My scope includes: [list items]. Based on these damage photos, are there any items I should add that aren't already in my scope?