Here is an uncomfortable observation about the remote audit your regulator just praised — it certified your paperwork, not your process.

SK Bioscience secured WHO GMP approval through what the industry called a streamlined remote audit. The headlines celebrated speed. Nobody asked what walked out of frame when the camera panned. I have been on both sides of AS9100 and EN9100 audits — as the auditee sweating through a tier-1 supplier review, and as the person asking the questions that make shift supervisors go quiet. The 50% reduction in EASA audit findings I delivered in one cycle did not come from polishing documents. It came from being on the floor at 06:00, catching what the procedure manual said versus what the operator at station 14 actually did with their hands. There is a gap between those two things. That gap is where your nonconformances live, and a webcam cannot find it.

What a camera doesn't see

A walking auditor catches things a screen-share will never find. Not one or two — entire categories of risk that remote formats structurally cannot address.

Operator body language. When I walk a line, I watch for the micro-hesitation. The half-second pause before an operator answers a question about a control plan. That pause tells me they are reconstructing the answer from training slides, not from muscle memory. On a video call, they look at a lens. They perform. The performance masks the gap between training and practice.

Environmental controls. A remote audit shows you a temperature log. It cannot show you whether the HVAC unit in the corner is actually running, or whether it has been rattling for three weeks because maintenance put it on "the list." I have walked cleanrooms that passed every documented check and failed the naked-eye test in thirty seconds.

Informal visual management. The best shops I have audited do not look like the 5S poster. They have tape on the floor that has been re-applied six times, handwritten notes on shadow boards, a whiteboard with yesterday's scrap count that nobody erased. That is real operational data. A camera pointed at a clean wall tells you nothing about how the place runs.

Undocumented workarounds are where it gets dangerous. Every operator in every plant I have walked has a little adjustment — a way of holding the part, a sequence tweak, a fixture mod — that is not in the work instruction. Some are latent genius. Some are latent defects waiting for a different operator on a different shift. You find them by watching hands move, not by reviewing PDFs on a shared screen. And then there is the smell of rework — solvent in the air where there should be none, the faint heat signature of a station used more than the traveller admits. You cannot stream that over Teams.

The audit is a sample, and the sample just got smaller

An audit is a statistical sample of reality. Nothing more. A good auditor with three days on site might observe 40 to 50 process moments across shifts, areas, and product lines. That is already a thin slice. Remote formats compress that sample by 30 to 40% in practice, even when the checklist says the scope has not changed.

The auditor can only see what someone carries the camera to. Document review eats more time over screen-share, so less time goes to process observation. Time zone differences cut productive hours. And there is a subtle but pervasive selection bias: the person holding the camera is usually the quality manager, who knows precisely which corners of the plant to avoid pointing the lens at. I am not describing fraud. I am describing human behaviour under audit pressure — the same instinct that makes a supplier pre-stage the cleanest parts for the incoming inspector. In person, a seasoned auditor can feel the staging and go look somewhere else. On a screen, you are guided through a rehearsed route and you do not even know what you skipped.

How to run a remote audit that isn't theatre

If you must go remote — and there are legitimate reasons for interim monitoring — four things separate an audit from a performance.

Require live body-cam gemba walks, not staged camera shots. The operator wears the camera. You see what they see, in real time, including the clutter, the distractions, and the moment they reach for a fixture that is not in the work instruction. If the quality manager insists on holding the device, you have already failed.

Pull documents unannounced during the walk. "Show me the PFMEA for the line you are standing on right now. Page 12 — the poka-yoke section for the torque station." If the answer is "let me get back to you," that is a finding, not a logistical issue. Sample across shifts, too. A remote audit that runs during business hours in one time zone is auditing the A-shift performance, not the management system. C-shift is where the real process lives, and I have found more nonconformances at 22:00 than at 14:00. Every single time.

Use two auditors. One on screen reviewing documents and interrogating the management system. One physically on the floor. This costs more than a pure remote audit and less than a full on-site audit. It is the only hybrid format that catches both layers — the paperwork and the practice — and it is the one nobody wants to pay for.

A remote audit is a surveillance tool, not a discovery tool. Use it to watch what you already know — never to find what you don't.

Key takeaways

  • Remote audits compress real audit scope by 30 to 40% — camera-guided observation cannot replicate unguided floor discovery, and supplier oversight models need to price that risk.
  • The blind spots of remote formats — body language, environmental controls, informal visual management, undocumented workarounds, sensory cues — are exactly where latent defects accumulate.
  • If remote is unavoidable, mandate body-cam walks, unannounced document pulls, cross-shift sampling, and always pair the screen auditor with boots on the floor.
  • Reserve remote audits for monitoring between physical audits. The cost of the blind spot will always exceed the cost of the flight.

The SK Bioscience approval will be cited as precedent. Other manufacturers will point to it and ask why they cannot do the same. Reasonable question. Uncomfortable answer. The gap between documented compliance and actual practice is widening right now, and the cost will surface in field failures, escape rates, and the audit finding nobody caught because the camera was pointed at the wall — not behind the cabinet where the expired calibration sticker was taped to a unit still in rotation. I have walked enough lines to know what a clean audit report is worth. Sometimes everything. Sometimes the PDF it was signed on. The difference is whether someone with sharp eyes and dirty boots was actually standing in the building.