Damage Assessment

The Damage Assessment pipeline analyzes your inspection photos to identify, classify, and document property damage, then generates a fully-costed repair scope. The AI works through six sequential steps, from raw image analysis through to validated line items with quantities and pricing.

---

Assessment Layout

Damage Assessments use a three-panel layout:

• Left panel — Image thumbnails organized by location group (persistent across all tabs)
• Center panel — Main workspace with four tabs
• Right panel — AI Chat assistant

Tab	Contents
Details	Uploaded images with damage bounding boxes, detected damage item cards (color-coded by severity), and Scope of Works line items
Measure	Drawing tools overlaid on the selected photo for annotating dimensions
Files	All uploaded images and documents
Team	Team members and access roles

---

Running the Pipeline

The pipeline processes all uploaded images automatically, end-to-end, without interruption. It is the fastest way to go from photos to a fully-costed repair scope. Use the Chat Agent to refine results, ask questions, or add items after the pipeline has run. For a full comparison, see Pipeline vs Chat.

1. Open a Damage Assessment with uploaded images.
2. Click the Run button.
3. The prompt dialog appears:
   • Optionally enter a custom prompt to guide the AI (e.g., "Focus on water damage in the ceiling area").
   • Leave the prompt empty to run general damage detection across all images.
4. Click Run to start.
5. A processing indicator shows real-time progress as each pipeline step completes.

The pipeline runs across all images in the assessment. Each image is analyzed individually, and results are consolidated into a single set of damage items and line items.

---

The 6-Step Pipeline

Step	Name	What Happens
1	Analyze Damage	Classifies all damage visible in images — sets damage type, bounding boxes (as percentage coordinates), severity, and description for each damage area
2	Generate Repair Scope	Creates repair line items (materials, labour, equipment) based on the damage analysis — no costs at this stage
3	Estimate Quantities	Resolves quantities for every line item: first checks documents, then drawings, then uses an engineering estimate — no item is left with quantity zero
4	Estimate Costs	Prices each item in mandatory order: Materials and Equipment first, then Other Costs (overhead, site services), then Labour last — every cost is sourced from a search result with an explanation
5	Calculate Totals	Computes group subtotals and the overall assessment total, then reports a final summary
6	Validate and Auto-Fix	Checks for unsplit groups, unpriced items, missing damage item references, and missing quantities — reruns the relevant step automatically if errors are found (single retry)

---

Damage Items

For each detected damage area, the AI creates a damage item containing:

Field	Description
Damage Type	Classification of the damage (see table below)
Damage Cause	How the damage occurred (Accidental, Wear and Tear, Gradual, Maintenance)
Severity	How serious the damage is (see severity table below)
Bounding Box	Rectangle on the image marking the exact damage location, expressed as [xmin, ymin, xmax, ymax] percentages of the image dimensions
Confidence Score	How certain the AI is about the detection (0–100%)
Policy Payout Type	Sudden (typically covered) or Gradual (may not be covered) — used in insurance assessments
Materials Affected	What materials are damaged
Fix Method	Recommended repair approach
Fix Area	Estimated area requiring repair (m²)

Damage Types

Type	Description
Abrasion	Surface scratching, scuffing, or wearing
Impact	Damage from physical force (dents, cracks, breaks)
Expansion/Contraction	Cracking or warping from temperature or moisture changes
Moisture/Water	Water stains, mold, rot, swelling
Temperature	Heat or cold damage (blistering, frost damage)
Chemical	Corrosion, discoloration, or degradation from chemicals
Fire	Burn marks, smoke damage, charring

Severity Levels

Severity	Meaning
Low	Minor damage, cosmetic or non-urgent
Medium	Moderate damage requiring attention
High	Significant damage needing prompt repair
Critical	Severe damage requiring immediate action

Severity is shown with color coding on the damage item cards and on the bounding boxes drawn over the images.

---

Reviewing Damage Items

After the pipeline completes, switch to the Details tab to review results:

1. Each damage item appears as a card showing damage type, severity (color-coded), affected materials, fix method, fix area, confidence score, and the linked image with its bounding box.
2. Click a damage item to highlight its bounding box on the image.
3. The Scope of Works section below the damage items shows the generated line items with quantities and costs.

---

Editing Damage Items

You can modify any detected damage item to correct or refine the AI's output:

1. Click on a damage item to select it.
2. Open the edit dialog.
3. Adjust any field:

Field	Description
Type	Change the damage type
Cause	Change the damage cause
Severity	Adjust the severity level
Materials Affected	Specify what materials are damaged
Fix Method	Describe how the damage should be repaired
Fix Area	Enter the area in square meters that needs repair
Location	Update the location description

4. Click Save to update the damage item.

---

Adding Damage Items Manually

If the AI missed something, you can add damage items by hand:

1. On the Details tab, click Add Damage Item.
2. Fill in the damage item fields (type, cause, severity, etc.).
3. Optionally draw a bounding box on the image to mark the location.
4. Click Save.

To remove a false positive, select the damage item, click Delete, and confirm. The damage item and its bounding box are removed.

---

Insurance Claim Context (Inoclaim)

When a Damage Assessment is run within an Inoclaim insurance assessment, steps 3 and 4 (Estimate Quantities and Estimate Costs) automatically apply additional insurance-specific rules:

Rule Area	Description
Full element measurement	The entire element is measured, not just the damaged area
Property age	Age of the property and affected elements is factored in
4-tier material identification	Materials are matched in order: exactly the same → almost the same → suitable replacement → like for like
Policy priority	Excess, limits, and betterment provisions are applied from the insurance policy
Replace vs repair analysis	Economic comparison to determine whether replacement or repair is more appropriate
Occupied property adjustments	Additional costs for working in an occupied property are included
P&G, travel, and return visits	Preliminaries, general conditions, travel costs, and return visit costs are included
Difficulty and access adjustments	Costs for difficult access or unusual site conditions are applied
By-law trigger costs	Where repairs trigger compliance requirements, by-law costs are included

These rules are loaded automatically — no additional configuration is required when working in Inoclaim.

---

Tips for Better Detection

• Good lighting: Well-lit photos produce better detection results.
• Multiple angles: Photograph damage from different angles to capture its full extent.
• Close-ups: Include close-up shots of damage areas for detailed analysis.
• Context shots: Include wider shots showing the damage in its surroundings.
• Custom prompts: Use specific prompts like "Focus on water damage on the ceiling" to direct the AI. You can also use the chat agent for follow-up analysis after the pipeline has run.

---

Next Steps

• Review and edit line items and costs
• Use the Chat Agent for deeper analysis, refinements, or report generation
• Export to Fergus, Xero, or CSV