Help Center → Analyze
AI Damage Detection — How Findings Work
How Estimatics AI analyzes your inspection photos to detect damage by type, severity, and location — and how to review, edit, and use AI findings in your workflow.
5 min read
AI Damage Detection — How Findings Work
When you upload inspection photos to a job, the Estimatics AI reviews every image and identifies damage automatically. Each piece of damage becomes a finding — a structured record of what was damaged, how severe it is, where it's located, and what material is affected. Findings are the foundation of your Analyze workflow: they feed into narratives, reports, and scope recommendations.
This article explains how AI damage detection works and how to get the most out of it.
What the AI detects
For each photo in your job, the AI evaluates:
- Damage type — hail impact, wind lift, water intrusion, cracking, missing material, biological growth, mechanical damage, and more
- Severity — classified as High, Moderate, or Low based on the extent and nature of the damage
- Location — the area or room the photo is assigned to (roof, exterior elevation, interior room, etc.)
- Material affected — the specific building component or surface (asphalt shingle, vinyl siding, drywall, flashing, etc.)
- Confidence level — how certain the AI is about the detection, expressed as a percentage
- Policy coverage assessment — a preliminary determination of whether the damage is likely Covered, Not Covered, or Review Required based on policy context
The AI also generates a repair recommendation for each finding, describing the likely remediation approach for the identified damage.
How detection works
- Upload your photos — capture photos in the field using the Estimatics iOS app, or upload them through the web dashboard. Assign each photo to an area for best results.
- AI processes in the background — once photos are synced and the job is ready, the AI analyzes every image. Processing typically takes a few minutes depending on the number of photos.
- Findings appear in the Analyze tab — open the job on the web dashboard, navigate to Analyze → Findings, and review what the AI found.
You do not need to trigger detection manually. The AI runs automatically when new photos are available and the job is in the appropriate state.
The Findings tab
Each finding is displayed as a card in the Findings tab. A finding card includes:
| Field | Description |
|---|---|
| Severity badge | Color-coded indicator — red for High, orange for Moderate, yellow for Low |
| Damage type | The category of damage detected (e.g., "Hail Impact," "Water Staining") |
| Source photo | Thumbnail of the photo that triggered the finding — tap to view full size |
| Location | The area where the damage was found |
| Material | The building component or surface affected |
| Confidence | Percentage score reflecting AI certainty |
| Coverage status | Whether the damage appears to be covered under the policy context provided |
| Repair recommendation | Suggested remediation for the damage |
Findings are sorted by severity by default, so the most critical damage appears first.
Reviewing and editing findings
AI findings are a starting point. You should review each one and make adjustments as needed:
- Verify accuracy — confirm the damage type and severity match what you observed in the field
- Adjust severity — if the AI rated something as Moderate but your professional assessment says High, change it
- Add context — use the notes field to add observations the AI cannot capture, such as cause of loss details or pre-existing conditions
- Dismiss false positives — if the AI flagged something that is not actual damage (a shadow, a pattern, normal wear), dismiss the finding
Your edits are saved and reflected in any reports or narratives generated from the findings.
Manual finding creation
On the Analyze+ tier, you can add your own findings alongside the AI-generated ones. This is useful when:
- You identified damage that the AI missed
- You want to document damage observed during a conversation with the homeowner
- You need to add findings for areas that were not photographed
Tap + Add Finding in the Findings tab, fill in the damage type, severity, location, and description, and save. Manual findings appear in the same list as AI findings and are included in narratives and reports.
Evidence quality and photo scores
The AI tracks an evidence quality score for each photo based on clarity, lighting, angle, and focus. Photos that score below the threshold are flagged so you can retake them while still on site. Higher-quality photos produce more accurate and detailed findings.
You can view photo scores in the Capture tab and in the finding detail view where the source photo is referenced.
Plan availability
AI Damage Detection is available on SCOPE and CERTIFY plans. FIELD plan users can upgrade to access AI findings.
Important note
AI findings supplement your professional judgment — they do not replace it. The AI is a tool to help you work faster and catch details you might miss, but the final assessment is always yours. Review every finding before including it in a deliverable.
AI Walkthrough — Guided Inspection Narration
Narrate your inspection while capturing photos — the AI transcribes and extracts findings automatically.
AI Narrative Report
Generate a professional narrative summary from your findings with one click.
Capturing Photos with the Mobile App
Photo technique tips for the best AI detection results.
Last updated: April 2026 · Questions? Use the Resources panel in the app or email support@aiestimatics.com
Last updated: April 2026 · Feedback on this article