
How multimodal AI surpasses traditional industrial sensors
Industrial sensors (accelerometers, telemetry) excel at monitoring machines but remain blind to manual operations: they detect a signal, not the context of what the workforce is doing.
The new edition of ISO 9001 is expected by autumn 2026 and is currently at the Final Draft stage (ISO/FDIS 9001, 2026). For the first time since 2015, the reference standard for quality management systems changes in both structure and content, and over the coming months the quality managers of thousands of manufacturers will open the transition project. Yet the problem that wastes the most time during an audit, and weighs most heavily on the margin, is not among the changes in this revision.
The changes confirmed by the standards and certification bodies shift attention toward the top of the organization. According to BSI (2026) and Bureau Veritas (2026), the new edition explicitly brings climate change into the organization's context, separates the management of risks from that of opportunities into two distinct requirements, and asks top management to actively promote a quality culture and ethical conduct. A new annex of application guidance is added, absent from the 2015 version.
These are changes of orientation, not of the shop floor. In line with the practice for management-system standards, the transition will presumably run up to three years from publication, with the exact deadline still to be confirmed in the IAF guidance. That is the window within which an already-certified company will have to align its system. The temptation, in these months, is to concentrate the entire effort on the new clauses; an understandable priority, but a partial one.
The requirement that generates the most audit findings does not change with the revision. It is the management of documented information, already present in ISO 9001:2015 at clause 7.5 and confirmed in the new edition (ISO/FDIS 9001, 2026). The standard requires operating procedures to exist, to be controlled and kept up to date. It cannot guarantee what really matters, namely that they describe the work as it is actually performed on the line. When the written procedure and shop-floor practice diverge, conformity becomes formal and the operational risk stays uncovered.

This gap between procedure and practice comes from three distinct mechanisms, worth keeping separate because each is fixed differently:
A documented procedure is not automatically a controlled procedure, and this is where the cost that no budget isolates opens up. A misaligned procedure produces false reassurance: the quality system marks the activity as covered, while on the line that activity still depends on the memory of a single operator. As a result the error does not disappear, it merely moves off the quality system's radar, and comes back as scrap, rework or a complaint, the cost-of-poor-quality items that erode the margin without concentrating in a single cost center.

A recurring example is the setup of a complex machine. The procedure describes the setup of a few years earlier; in the meantime a fixture has changed and the experienced operator has adapted the sequence without updating the document. As long as he is there, the line runs. When a new hire comes in and follows the procedure exactly as written, the part comes out of tolerance, and the cost of the misalignment shows up all at once.
Not all procedures, however, weigh the same. The gap is tolerable on a marginal instruction, and becomes critical on the steps that determine safety, product conformity or scrap. That is where alignment is worth concentrating, and that is where a serious audit looks. Conversely, the company that keeps procedure and practice together cuts two costs in parallel, the time spent reconstructing the correct version during an audit or a handover, and the variability between those who learned from the document and those who learned from a senior colleague. In multi-site companies, where the same operation must give the same result in different plants, that gap is among the first causes of performance gaps between plants.
The useful conclusion, for those preparing the transition to ISO 9001:2026, is not to chase only the new clauses but to close the gap that the standard alone cannot fill. A procedure should be treated as a living object, aligned with the real process and updated when the process changes, not as a document to show the auditor. It is a change of function before it is a change of tool, and for that reason it is not solved by writing longer manuals.
On this ground technology has a concrete role. Some manufacturers are tackling the problem by digitizing shop-floor knowledge at the source: the video of an operation becomes a structured, versioned operating procedure that starts from how the work is actually done rather than from how someone imagines it should be done. Documentation thus stops being a formality and goes back to being what the standard intended, the operating memory of the company.
In a sector where experts are ever harder to find, keeping what they know in transferable form is not one audit requirement among others. It is how a company decides not to lose its own knowledge every time someone walks out the gate.

Industrial sensors (accelerometers, telemetry) excel at monitoring machines but remain blind to manual operations: they detect a signal, not the context of what the workforce is doing.

The disruption of flows in the Strait of Hormuz dictates a drastic revision of procurement strategies for Italian manufacturing.

The departure of senior technicians leads to a loss of tribal knowledge that increases Mean Time to Repair by 40-60%, directly eroding EBITDA.