Ask where inspection data disappears and people picture a lost notebook or a corrupted file. The more common answer is quieter: it evaporates at shift change, in the ten minutes between a tired day-shift technician and a night-shift crew that was not there when the finding was made.
During a turnaround this handoff happens twice a day for weeks, across multiple crews and contractors. Every verbal-only handoff is a lossy copy. By the third shift, "hold point on the east nozzle, do not close until the engineer sees it" has become "something about a nozzle."
What actually gets lost
- Partial work state. A UT grid two-thirds shot, a vessel internal half examined. If the incoming technician cannot see exactly which CMLs are done, they either redo work or, worse, assume completion.
- Unfiled findings. The day shift saw something an hour before quitting time and planned to write it up tomorrow. Tomorrow they are on a different unit.
- Pending holds and callbacks. The wait-for-the-engineer instructions that exist only as a verbal flag. When the engineer arrives at 21:00, nobody on nights knows why.
- Context behind decisions. Day shift skipped three CMLs because insulation was not stripped. Without that note, nights reports them as refused access, planning reads it as scope growth, and a small delay becomes a meeting.
The handoff as a system property
A reliable handoff is not a better briefing template. It is a property of how work is captured during the shift. If every reading, finding, and hold lands in the record at the moment it happens, then shift change transfers a live state instead of a memory.
The mechanics that make this real:
- Status lives at the task level. The work pack decomposes to items small enough that "in progress" is unambiguous: this CML shot, this weld examined, this internal zone complete. The incoming crew opens the pack and sees the actual frontier of the work.
- Findings are filed before boots leave the unit. Capture at the point of observation, offline if the unit has no coverage, synced by end of shift. The rule worth enforcing: a finding that is not in the system did not happen.
- Holds are records with owners. A hold point carries who set it, why, and who must release it. It blocks the dependent task in the system, so nights cannot close the nozzle by accident even if the briefing missed it.
- The briefing is generated, not composed. The outgoing lead reviews a system-produced summary of the shift: completed items, new findings, open holds, exceptions. The conversation becomes verification and nuance, which is what conversations are good at.
The contractor wrinkle
Handoffs fail hardest across company boundaries. Your day-shift NDE contractor and your night-shift crew may not share a trailer, let alone a database. If contractors capture into their own systems and deliver reports at demobilization, every shift change during the job is a blind handoff. The fix is contractual and technical: field crews work in the owner's system, or their data syncs into it at shift cadence, not at job end.
A fair test of your current handoff: at 02:00, can the night lead determine which CMLs on a given drum remain unshot, without calling anyone? If the answer requires a phone call to someone asleep, the state lives in people instead of records.
intelliSPEC's field capture, offline sync, and crew assignments in intelliCREW are built so the record carries the shift state and the handoff reads from it. Walk a real shift change through it at a demo.
