A short walk becomes a report full of findings.
No special hardware. No new routines. No change to who walks or when. Here's exactly how, step by step.
Five steps.
No scanning rig, no site visit, no change to the walk.
Here's exactly what happens after your team hits record.
Your team films the walk
A routine walkdown on a phone or GoPro. Anywhere from 45 seconds to a couple of minutes is plenty. No special protocol, no new training, no kit to carry.
Footage uploads through your browser
Secure, encrypted, in-browser. No IT involvement, no site access from us, no scanning equipment. The walk goes straight up and it's on its way.
duallo rebuilds the walk and reviews every frame
The footage becomes a navigable visual record, and every frame is checked for leaks, corrosion and hazards. The things easy to walk straight past.
A structured findings report lands
Each finding carries the moment in the walk it was seen and visual context, ready to route into your CMMS or work-order flow.
You review the finished model, every finding pinned in place
This is the actual output from the 50-second walk above. Spin the model, zoom in, and tap any pin to jump to the exact frame it came from.
Reconstructed from 50 seconds of walkdown footage · drag to rotate, scroll to zoom
.ply at assets/model.plyThat's the whole pipeline. Send one walk and get your own findings on a model like this.
Computer vision flags. People verify. Your team decides.
duallo isn't a black box that emails you raw AI output. Every finding passes a person before it reaches you. That's the part that makes it usable.
Computer vision flags
Models trained on industrial walkdown footage watch every frame and surface potential issues: leaks, corrosion, CUI risk markers, missed inspection points.
A reviewer verifies every flag
Before anything reaches your team, a trained reviewer checks each flagged candidate and clears the noise, so you're never handed raw model output to sort through.
Your team decides
Findings arrive as leads, not conclusions. What matters and what gets actioned stays with your inspectors and maintenance team.
That human-in-the-loop step is why we call the output leads, not conclusions. The model catches what's easy to miss, a person confirms it's real, and the judgement stays with you.
It plugs into the work your team already does.
No new routine to adopt. duallo feeds the cycles and systems you already run.
Routine walkdowns
Your team films the walks they already do. No change to routes, schedules, or who walks.
MI inspection prep
Visual evidence of plant state between formal API 510 / 570 / 653 cycles.
PHA, MoC & RCM
What's actually on the plant right now, not what was on last cycle's drawings.
CMMS work orders
Findings export in a structure that flows into your existing system.
If you've been pitched LiDAR or digital twins, this isn't that.
duallo is a findings tool, not a scanning or documentation tool. The visual record is the mechanism, findings are the product.
- Not a 3D scan, point cloud, or digital twin.duallo doesn't replace LiDAR vendors or scanning services.
- Not a documentation system.Doesn't update P&IDs, PFDs, or schematics.
- Not a replacement for certified MI inspections.duallo lives between routine walkdowns and formal MI cycles, not in the place of them.
- No specialised hardware.A phone or GoPro is enough. No drones, no scanners, no scheduled capture missions.
That's the process. The interesting part is what it finds.
Send a short walkthrough from a single unit and get your own findings back. No commitment, and the outputs are yours to keep.