IN Brief:
- Thirty-nine per cent of UK manufacturers expect operational efficiency to become the main future value of security systems.
- Only 18% currently use security technology for productivity, while 16% fully integrate its data into operational decisions.
- Food and beverage manufacturers identify compliance, quality rework, and limited real-time visibility as major areas of pressure.
Axis Communications has identified a wide gap between manufacturers’ interest in visual intelligence and the use of video data within day-to-day operational decisions.
A survey of 600 UK manufacturing decision-makers found that 39% expect operational efficiency and productivity to become the greatest value delivered by security systems during the next three to five years. Security and incident prevention ranked second at 33%.
Current use remains concentrated on conventional protection, with 96% applying security technology to security or incident prevention while only 18% use it to support operational efficiency and productivity.
Although 45% see significant or some potential for network cameras and visual data beyond physical security, only 16% said the resulting information was fully integrated into operational decision-making. A further 42% use it in some decisions, while 27% do not use it operationally.
Rising costs were identified as a major challenge by 57% of manufacturers, and 26% selected operating-cost reduction as the area where improved insight could have the greatest immediate impact.
Among food and beverage respondents, 61% cited regulation and compliance as a major pressure. Forty-one per cent expect efficiency and productivity to become the leading future use of security systems, while 20% prioritised limited real-time visibility and 22% selected quality defects or rework.
Visual records add context to process data
Food factories already generate information from programmable controllers, sensors, inspection equipment, laboratory systems, production software, and maintenance platforms. Those records can show when a stop occurred or a reject rate increased without necessarily revealing the physical event that caused the change.
Video can connect a timestamped process deviation with a product jam, operator intervention, unstable package, obstruction, spill, or cleaning activity that produced no independent machine signal.
Synchronising footage with production records allows teams to investigate transient events without relying entirely on recollection or attempting to recreate a fault that has disappeared. The same approach can help trace a quality defect backwards when the visible failure appears downstream from the point where it originated.
Defined analytics can monitor whether protective equipment is worn, restricted areas remain clear, accumulation is developing, or guards and doors are in the expected position. Visual or audible alerts may then prompt intervention before a condition develops into an incident.
Such systems need careful validation because lighting, clothing, steam, washdown, reflections, dust, and changing line layouts can affect image analysis. An application performing reliably in a warehouse may require substantial adaptation before it can operate within a food-processing hall.
Cameras and analytics should also complement rather than replace physical safeguards, interlocks, supervision, and safe process design. A visual alert can draw attention to a problem, but it cannot perform the protective function of a correctly engineered guard.
Governance follows operational integration
Moving video from a security control room into production expands the number of people and systems using the data. Access rights, employee privacy, retention periods, cyber security, and permitted uses must be defined before footage becomes part of routine process management.
Trust will depend on clarity over whether a system is monitoring a machine condition, investigating a production event, or assessing individual behaviour. Poorly defined use can create resistance even where the technical application is legitimate.
Data volume presents another constraint because continuous high-resolution footage from a large factory requires substantial network and storage capacity. Edge analytics can reduce the burden by processing images close to the camera and transmitting only events or selected recordings.
Operational integration must remain selective. Sending every camera notification into a maintenance or production platform would create noise, so events need priorities, ownership, thresholds, and a defined response.
Only 14% of respondents currently use security technology to improve health and safety, while 37% of health and safety specialists were unsure of, or had not considered, applications beyond conventional security.
PPE monitoring, restricted-area controls, and detection of unsafe behaviour can support earlier action, although the systems require testing against false positives and missed events. A model trained around one uniform, product format, or camera angle may not remain reliable after routine factory changes.
Linn Storäng, regional director for Northern and Eastern Europe at Axis Communications, said: “The smart factory will not be built on data alone. Manufacturers also need context, and network camera technology can provide a vital layer of visual intelligence that many organisations have not yet fully explored.”
Inspection and production systems are already moving towards more connected operating environments. The integration of inspection, grading, and automation within poultry-processing systems demonstrates how data from individual machines can support a broader view of quality and line performance.
Video can extend that visibility to physical events not captured by conventional sensors, provided each application begins with a defined operational problem. Additional cameras without clear decisions, responsibilities, or performance measures will produce another isolated data source.
Initial projects are likely to concentrate on persistent downtime, rework, safety, and compliance rather than converting every security camera into a production sensor. These areas provide recurring events against which reductions in investigation time, waste, or incidents can be measured.
Visual intelligence will become useful when it shortens root-cause analysis, prevents repeated faults, and prompts timely intervention. The technology already records much of what occurs on the factory floor; the remaining task is to connect those images with controlled operating decisions.
