Help Center → Capture
Organizing Media by Area
Area tags are how Estimatics knows where each photo was taken. The 100-entry room catalog, the AI that suggests a tag, and the one rule that decides which photos it will guess on.
5 min read
Organizing Media by Area
Every photo and video belongs to an area — the label that tells Estimatics, and the AI, where on the property it was taken. This is the single highest-leverage habit in the whole workflow: a photo that belongs to "Master Bathroom" produces a materially better finding than a loose photo, because the model knows what it's looking at.
Organization also drives scope accuracy (line items are grouped by area), report structure (reports read area by area), and geometry (dimensions are per-area).
The room catalog
Estimatics ships 100 pre-seeded room types, grouped into three categories and fifteen subcategories:
| Category | Entries | Covers |
|---|---|---|
| Residential | 52 | Living areas, bedrooms, bathrooms, kitchen & dining, garage & storage, exterior, roof |
| Commercial | 38 | Common areas, work areas, retail, commercial kitchen, utilities, elevation, exterior, roof |
| Standard | 10 | General-purpose areas that apply to any property |
Each entry carries a color, assigned by subcategory — bedrooms indigo, bathrooms cyan, kitchen and dining red, exterior emerald, garage and storage gray — so the gallery reads at a glance.
Exterior (18 entries) and Roof (13) are the two largest subcategories. That's deliberate: they're where most claims are argued.
Custom areas
If the catalog doesn't match the property, add your own.
On iOS during capture: tap the area selector, scroll to the bottom, add the area, and it's available immediately.
On the web: Capture tab → Manage Areas → add.
Custom areas sync both directions immediately and behave identically to catalog entries — same filtering, same report treatment.
For commercial properties, create your areas before the inspection starts — "Building A", "Unit 101", "Building C Roof". Setting up structure beforehand is the difference between an organized file and 200 untagged photos.
The AI tag suggestion
When an overview photo uploads, Estimatics analyzes it against the room catalog and proposes a tag with a numeric confidence. You accept it or you don't — a suggestion sits as pending until somebody decides.
Three states: pending (proposed, awaiting you), accepted, rejected.
If the model isn't confident enough to name a catalog entry, it returns nothing rather than guessing.
It only guesses on overview photos
This is the part worth understanding, because it explains why most of your photos never get a suggestion.
Every photo is first classified by intent — what kind of shot it is:
| Intent | What it is |
|---|---|
| Overview | The wide shot that establishes a room or elevation |
| Damage detail | The close-up of the damage itself |
| Element | A specific component — a fixture, a vent, a joint |
| Work progress | Work underway |
| Document | A photographed document |
| Context | Surrounding context |
Only overview photos get a room suggestion. A close-up of a water stain can't identify which room it's in — asking a model to guess from one is how photos end up in the wrong area. So it doesn't ask.
This is why the wide shot matters. Shoot the overview first and the rest of that room's photos inherit an area that was set correctly once.
It never overrides you
The classifier stands down completely when:
- You tagged the photo manually — a human decision is final, including a deliberate un-tagging
- The tag came from a LiDAR scan, which assigns areas deterministically
- The suggestion was already accepted or rejected
Your tag is never silently replaced.
Tagging manually
- During capture — pick the area before you shoot. Fastest workflow by a distance, and it removes the sorting pass entirely
- After capture — open the photo and change the area
- In bulk — select multiple items in the gallery and reassign them in one action
Managing tags
From the job's area panel: rename (assignments are preserved), change color, reorder (the order you set is the order used in the report), delete (photos are untagged, not deleted), and read the photo count per area.
Tag changes broadcast live — an inspector tagging exterior photos in the field updates what a reviewer sees on the dashboard without a refresh.
Untagged media
Media with no area is grouped under Unassigned. Get that to zero before you consider the inspection complete.
Filter to Unassigned, assign each item, repeat. An area with zero photos is the other signal worth reading: it means you didn't document it.
Frequently asked questions
Can I rename an area after photos are tagged to it? Yes. Every assignment follows the rename.
Can I delete an area? Only when it holds no media. Reassign first, then delete.
Can I reorder areas? Yes, by dragging in Manage Areas. That order carries into the report.
Is there a limit on custom areas? No.
Do area tags actually change the AI findings? Yes. Area context is part of what the analysis reads. Untagged or miscategorized photos produce weaker findings.
Why did this photo not get a suggestion? Almost always because it isn't an overview shot. See above.
Capturing photos
Technique that produces accurate findings and good photo scores.
Media gallery
Browsing, filtering, and managing the full media library.
How media syncs
How tags and media move between iOS and the dashboard.
Photo markup
Annotating the photos you've organized.
Last updated: September 2026 · Questions? Email support@aiestimatics.com
Last updated: September 2026 · Feedback on this article