INSPWORKSExplore the demo

RESEARCH & EVIDENCE

A clear direction.
A testable starting point.

We are studying inspection software, reality capture and industrial data systems to understand where INSP Works can make the work simpler. Here is what informs the product, and what still needs proof.

THE CENTRAL QUESTION

Can we reduce the cost
of a complete, accepted job?

Drawing preparation is a promising starting hypothesis. We have not established that it is the largest cost across inspection programs.

A faster capture or a more convincing model is only part of the result. Records, access, reconciliation, drawing corrections, review, repeat visits and client acceptance all consume effort.

Our proposed first pilot follows one defined area from the existing information through to agreed drawings and reports. It compares equivalent deliverables and includes processing, equipment, review and rework.

Build a complete cost comparison

WHAT THE MARKET ESTABLISHES

The pieces already exist.
The connections matter.

01 / INSPECTION SYSTEMS

History and review are core product functions.

Cenosco documents equipment condition history, schedules and approval states. Antea describes inspection records with 2D, 3D and point-cloud context. These are established product categories, not new inventions.

Cenosco PEI documentation · Antea IDMS

Our implication: concentrate on a useful, traceable journey between capture, asset evidence and the delivered report.

02 / REALITY CAPTURE

Reconstruction is a pipeline, with quality checks.

COLMAP describes recovering camera poses and sparse structure before dense reconstruction. Bentley supports image and point-cloud processing with control points and quality tools.

COLMAP reconstruction tutorial · Bentley iTwin Capture Modeler

Our implication: compare established reconstruction tools before committing to a new engine. Test dimensions and coverage independently.

03 / INDUSTRIAL CONTEXT

A visual model becomes useful through its records.

Cognite describes connecting industrial documents, drawings, models and operational data. GE Vernova presents role- and task-oriented demonstrations across its asset performance portfolio.

Cognite industrial digital twin · GE Vernova demo hub

Our implication: let people follow a specific asset and task, with links to the information behind the screen.

04 / AI ASSISTANCE

Evidence and authority need visible boundaries.

NIST's AI Risk Management Framework addresses context, measurement, ongoing risk management and human oversight. It is guidance, not a certification for this product.

NIST AI RMF core

Our implication: show sources, preserve uncertainty and separate draft assistance from an authorized inspection decision.

CURRENT EVIDENCE

Know what you are looking at.

The public demonstration is a small, working example of the intended experience.

Working in this website

Select fictional assets, read sample evidence, edit a draft, save page-session revisions, download the draft and compare entered pilot costs.

Being developed

Field capture, source reconciliation, drawing preparation and review mechanics are being investigated through local feasibility work.

Still to be established

Field accuracy, client acceptance, measured savings, production integrations and a live AI assistant need further development and validation.

THE NEXT EVIDENCE WE NEED

Five questions for a field pilot.

  1. Which deliverables consume the most total effort for this inspection program?
  2. What capture method covers the required components under the actual site conditions?
  3. How accurately can the proposed drawing locate and describe the required inspection points?
  4. How much correction and professional review does the assisted result require?
  5. Does the accepted result cost less after all capture, processing, review and delivery costs are included?

Answers may differ across corrosion surveys, turnarounds and recurring equipment inspections. A result in one setting should not automatically become a claim for every setting.

How this research is used.

These links are primary sources reviewed on 4 October 2026. Vendor documentation establishes what a vendor describes; it does not independently verify performance or predict our results. Product implications on this page are our interpretation. No affiliation or endorsement is implied.

We also use the W3C guidance on reflow and target size to inform the website's layout and controls. This is not a claim of formal accessibility conformance.

Explore the working example

FROM CAPTURE TO DRAWING

Ask what each view actually establishes.

COLMAP describes reconstruction from overlapping photographs taken at different viewpoints and cautions about weak texture, reflections and camera rotation without movement. Its guidance informs how we think about capture coverage. COLMAP capture guidance.

Our product interpretation is to keep visible evidence, proposed connections and dimension review separate. The public walkthrough illustrates that distinction: an added view reveals a branch while a placement disagreement remains open. These are fictional examples, not published field-performance results.

Explore the capture walkthrough