How EnvelopX turns a single drone flight into a structured, AI-powered building intelligence report
10x
Faster than manual facade inspection
99%
Anomaly detection accuracy, validated on real commercial building data
235+
Camera positions captured per inspection run
Zero
Scaffolding, rope access, or cherry-pickers needed
Industry
PropTech / AI & Computer Vision / Drone Technology
Organization
EnvelopX (formerly Project Tracer POC), an AI-powered commercial building inspection platform built by Attri and listed in the Twelvefold Venture portfolio. Targeting property managers, construction firms, insurance assessors, and government infrastructure teams.
Challenge
Traditional building inspections require scaffolding or rope access, produce inconsistent documentation, and cannot track anomaly trends over time. No platform existed to convert drone imagery into structured intelligence.
Solution
A full-stack AI drone inspection platform covering automated flight planning, photogrammetric 3D reconstruction, computer vision anomaly detection, side-by-side review, severity classification, and repair tracking.
The Challenge
Commercial building inspections are slow, expensive, and dangerous. Facades, rooftops, and hard-to-reach structural areas require scaffolding, rope access, or cherry-pickers to inspect, adding weeks to timelines and tens of thousands in access costs. Even when data is collected, findings are documented inconsistently across spreadsheets and photos, making it nearly impossible to track whether a crack is growing, assign repair accountability, or produce an audit-ready report. No platform existed to turn drone imagery into structured, trackable building intelligence.
Traditional inspections require scaffolding, rope access, or cherry-pickers
Physical access to facades, parapets, and hard-to-reach areas adds weeks to inspection timelines, significant cost, and real safety risk for personnel.
Inspection findings documented inconsistently
Reports were produced in different formats by different inspectors (spreadsheets, photo dumps, narrative PDFs) with no structured data model to enable cross-building or cross-time comparison.
No structured way to track anomaly trends over time
Whether a crack identified in last year's inspection had grown, stabilised, or been repaired was impossible to determine without manually cross-referencing unstructured reports.
Drone flight planning required manual waypoint setup
Setting up a drone flight for a complex commercial building required manually placing hundreds of waypoints, a specialised task that added time and limited who could run an inspection.
No platform to turn drone imagery into actionable intelligence
Even firms using drones had no system to process imagery into structured anomaly data, assign confidence scores, triage by severity, and produce an audit-ready report. It was still a manual process from images to insights.
The Solution
Attri built EnvelopX, a full-stack AI drone inspection platform that covers the complete lifecycle from data capture to repair tracking. One drone flight generates a photorealistic 3D model of the building, every anomaly is pinned to it with a confidence score, and the reviewing team works through a structured triage interface, with no scaffolding, no manual documentation, and no guesswork.
Automated Drone Flight Path Planning
Computes optimised flight paths from building geometry, capturing 235+ camera positions in a single run. No manual waypoints required, so any operator can run a complete inspection without specialist setup expertise.
Photogrammetric 3D Model Reconstruction
Drone imagery is processed into a photorealistic 3D model of the building that serves as the inspection canvas, so anomalies are pinned directly onto the model rather than buried in a photo folder.
AI Anomaly Detection
Computer vision models scan every image and pin detected anomalies (structural cracks, concrete spalling, moisture intrusion) directly onto the 3D model with confidence scores. Image slicing is optimised automatically from drone EXIF metadata.
Side-by-Side Anomaly Review Panel
Each anomaly shows the original drone image alongside the AI-highlighted version, with a confidence score and four action buttons: Repair, Defer, Ignore, or False Positive. Triage is fast, structured, and documented.
Anomaly History & Repair Tracking Module
Trend charts across inspection runs, a sortable anomaly table, repair cost tracking with GL code support, vendor assignment, and Excel or PDF export, turning historical data into a living asset management system.
Real-Time Dashboard
Live breakdowns by Anomaly Type, Severity (Critical, Moderate, Low), and Status in a single view, giving property managers and inspectors an instant read on building condition across the portfolio.
Additional Features
Automated Flight Planning
Building geometry becomes an optimised flight path with 235+ camera positions and no manual waypoints.
Drone Data Capture
A single drone run captures the full building facade, rooftop, and structural surfaces at high resolution.
3D Model Reconstruction
Photogrammetric processing converts drone imagery into a photorealistic 3D building model, the inspection canvas.
AI Anomaly Detection
Computer vision scans every image and pins anomalies to the 3D model with confidence scores and severity classification.
Reviewer Triage
A side-by-side review panel pairs the original image with the AI highlight, and the reviewer marks each anomaly as Repair, Defer, Ignore, or False Positive.
Report & Tracking
Audit-ready reports export to Excel or PDF, with repair costs, GL codes, vendor assignments, and trend history stored for future runs.
Anomaly Types Detected
Structural Cracks
Surface fractures across concrete, masonry, and cladding, tracked by location and severity over time.
Concrete Spalling
Delamination and surface loss on concrete facades, flagged before deterioration reaches reinforcing steel.
Moisture Intrusion
Water ingress and saturation indicators, detectable with standard imaging and thermal drone scans.
Works with your existing tools
The Results
| Metric | Before | Now |
|---|---|---|
| Inspection speed | Weeks of scheduling, access rigging, and manual inspection | 10x faster, a single drone flight replaces weeks of physical access |
| Anomaly detection accuracy | Dependent on inspector expertise and physical access angle | 99%, validated on real commercial building data |
| Camera positions per run | Limited by physical access points, coverage gaps common | 235+ camera positions, automated and complete coverage in one run |
| Scaffolding required | Always, with significant cost and scheduling overhead | None, the drone replaces all physical access equipment |
| Anomaly documentation | Inconsistent, spread across spreadsheets, photo dumps, and narrative PDFs | Structured, pinned to the 3D model with confidence scores and severity levels |
| Trend tracking across runs | Manual cross-referencing of unstructured reports | Automated history module with trend charts, cost tracking, and vendor assignments |
| Repair accountability | No structured assignment or GL code tracking | Vendor assignment, GL codes, and repair cost tracking per anomaly |
One automated drone flight generates a complete, structured, AI-annotated building inspection report with no scaffolding, no manual documentation, and no specialist drone operator.
Anomalies are spatially located on a photorealistic 3D model and visualised in context rather than buried in a folder of unstructured images.
The platform was validated on real commercial building inspections, including thermal drone scans, so the 99% detection accuracy reflects real-world performance.
Each inspection run compounds building history through trend charts, anomaly progression tracking, and repair records, turning EnvelopX into a living asset management system.
Why It Worked
Flight planning is fully automated
The system computes optimised flight paths from building geometry, with no manual waypoints and no specialist operator knowledge required, so any team can run a complete inspection.
Image slicing optimised from drone metadata
The system determines correct image splits from drone distance metadata (EXIF data) and resolution automatically, with no manual per-image decisions. It is a core part of the technical moat.
Anomalies pinned to a 3D model, not a photo folder
The photorealistic 3D building model is the inspection canvas, so anomalies are spatially located and visualised in context rather than buried in a folder of unstructured images.
Validated on real commercial buildings
The platform has been tested on real commercial building inspections, including thermal drone scans, so the 99% detection accuracy reflects real-world performance, not controlled lab conditions.
Built for the people who act on inspection data
Property managers, insurance assessors, and repair coordinators need actionable outputs, not raw data. The triage interface, severity classification, vendor assignment, and GL tracking are all designed for the people downstream of the inspection.
Historical intelligence compounds over time
Each inspection run adds to the building's history. Trend charts, anomaly progression tracking, and repair records turn EnvelopX into a living asset management system, not just a point-in-time report.
Security & Compliance
Get in touch
See how EnvelopX can turn your drone imagery into structured, audit-ready building intelligence.
Ready to automate
Get in touch
See how Attri can automate your document workflows in minutes.