Inspection software rollouts rarely fail in the office. They fail at the vessel, when an inspector in gloves, on a scaffold, in glare, decides the app is slower than the clipboard and quietly goes back to paper. Training determines which way that moment goes, and the standard approach, a conference room session two weeks before go-live, is close to worthless for field crews.
Why classroom training does not transfer
A slide deck teaches where the buttons are. It does not teach what the job actually requires: completing a CML route with cold fingers, recovering when a reading looks wrong, trusting that data captured offline will still exist after sync. Those are learned at the equipment, under field conditions, with someone nearby who has done it. Two weeks of gap between training and first use erases most of what the classroom delivered anyway.
What works instead
- Train on the crew's own route. Load the pilot unit's real equipment, real CMLs, real history. An inspector who sees last year's reading for the exact nozzle in front of them understands the system's point in one visit. Generic demo data teaches that the system is generic.
- Do the first route paired. The first live route for each inspector happens alongside someone fluent: a vendor trainer, or better, a peer from the pilot crew. One accompanied route beats three classroom hours.
- Build champions before the rollout. Two or three respected field hands from the pilot unit, involved early enough to have complained and been heard, carry more adoption weight than any manager. Their names on the rollout matter. Pay for their time explicitly; champion work is work.
- Rehearse the failure cases. Confidence in offline capture is the whole game for remote units. Have each inspector deliberately work in airplane mode, complete a route, and watch it sync. An inspector who has seen recovery work once will trust it. One who has not will keep a paper backup forever.
- Keep sessions short and role-shaped. Inspectors need capture, review, and history lookup. Coordinators need scheduling and reports. Chief inspectors need approvals and the overdue list. Ninety minutes per role, hands-on, beats a half-day tour of every menu.
The paper question
Decide, explicitly, when parallel paper ends. Running paper and app together "for safety" feels prudent and destroys adoption: the paper stays primary because it is familiar, the app becomes double entry, and double entry breeds resentment and errors in both places.
The workable pattern is a short defined overlap, two to four weeks, with a named end date announced at training. After that date, the app is the record. Leadership has to hold this line the first time a report is late, because that is the moment the organization learns whether the change was real.
Measure adoption, not attendance
Training completion is not the metric. Watch instead:
- Routes completed in-app as a share of routes performed, weekly.
- Sync latency: time from field capture to visible record. Rising latency means batching at a desk, which means the field workflow is not working.
- Rejected or corrected readings, which reveal where guidance is unclear rather than where inspectors are careless.
- Voluntary lookups: inspectors checking history before a visit without being told to. This is the signal that the system has become a tool rather than an obligation.
Numbers stalling in week three point at friction to remove, not at retraining to schedule.
We built intelliSPEC's field app assuming gloves, glare, and no signal, but no interface eliminates the need for this kind of rollout. Budget training as seriously as migration, and put it at the equipment, not in the conference room.
