Connected operations begin where visibility ends
Manufacturers have spent years making the factory visible. Machines are connected, production data is collected and dashboards show performance in detail. Yet the decisions that follow often remain fragmented, with production, maintenance, quality and supply chain teams still acting from different systems, priorities and measures of success.
For Ben Linke, President of Industrial Automation for Europe at Honeywell Technologies, that gap explains why many digital programs have delivered more information without creating a more coordinated operation. The next stage is not another layer of reporting, but an environment in which data moves into workflows, decisions and interventions quickly enough to change what happens on the factory floor. “The challenge is not really the lack of data,” Linke says. “It is that the information is spread across different systems and teams, and each group has its own priorities, focus areas and KPIs. IT environments have become more harmonized over time, but the operational side is often still a multitude of systems and machinery acquired over decades.”
Visibility was only the first step
Industry 4.0 encouraged manufacturers to connect assets that had previously operated in isolation. That work created a digital representation of production, but it did not automatically create a common operating model. Maintenance may see asset condition, quality may monitor compliance and production may track output, while none has enough context to understand the effect of its decisions on the wider operation.
The problem becomes more pronounced across multi-site manufacturers because every facility has accumulated its own equipment, software and practices. Plants producing similar products may use different historians, control systems or definitions for the same indicator. Connecting the data is possible but making it comparable and useful across functions requires industrial context as well as technical integration.
“Production talks about output, quality looks at compliance and maintenance focuses on keeping the assets running,” Linke explains. “They are all using different data sets and optimizing their own part of the value chain. The real objective is to bring that operational information together as part of one process, then use it to improve the performance of the whole facility.”
That distinction separates connected operations from a factory with more sensors. Connectivity provides access to information; connected operations determine how that information changes a decision. The architecture must extract data from legacy and modern systems, combine local and cloud-based sources where appropriate, and apply analytics or AI to identify the action that should follow.
The operating context matters more than the dashboard
A truly connected operation does not ask an operator to interpret an expanding wall of charts. It provides enough context to recognize a developing problem, understand its operational consequence and initiate the appropriate response. Sometimes that response will remain human-led; in other cases, the system can intervene automatically within defined limits.
Honeywell Technologies’ plant in Lotte, Germany, offers a practical illustration. The facility uses vertically integrated digital systems to support a high degree of product variety without repeated manual adjustments. Semi-automated error-proofing workstations guide employees through assembly, detect an incorrect component and stop the process before the mistake moves further down the line. The same environment reconfigures production systems when products or orders change, reducing manual intervention. Camera-based tracking analyzes vehicle movements within the facility, allowing the logistics team to improve routing, reduce waste and create a foundation for autonomous mobile applications.
“This is where connectivity turns into an operating system rather than a reporting system,” Linke says. “The workstation is not simply telling the employee that something is wrong after the event. It identifies the incorrect part, alerts the person and stops the process. The digital system also allows people with less experience to work with knowledge built up over many years.”
Linke does not disclose detailed financial results but says the facility has achieved almost no production failures, very strong first-time yield and one of Honeywell Technologies’ best on-time delivery performances. Working capital is also lower than at many comparable sites because internal logistics have been optimized around material and order flows. These outcomes show why the value of connected operations is cumulative. A workstation that prevents one error improves quality, but the same information can influence training, maintenance and scheduling. Once operational events are connected to wider consequences, the factory can improve the process rather than repeatedly treating symptoms.
The hardest silos are organizational
Manufacturers often describe legacy technology as the main barrier to connected operations, but integration capabilities have improved considerably. Modern industrial platforms can collect data from historians, enterprise systems and older equipment without requiring complete replacement. The larger obstacle is deciding how teams will use the information once it becomes visible.
Functional targets can work against a shared outcome. Maintenance may prefer to stop equipment before a component deteriorates, while production resists interruption during a demanding schedule. Quality may identify a trend that remains within specification, while supply chain is already committing output to a customer. Better data makes these tensions clearer but does not resolve them automatically.
Linke has seen manufacturers respond by bringing functions together around the same evidence rather than asking each department to interpret separate reports. “Once the data is visible, it challenges teams to think differently,” he says. “We see people coming together cross-functionally to problem-solve and decide how to continue improving. The pressures around productivity and skilled labor are forcing companies to move beyond their traditional swim lanes.”
Connected operations therefore require changes in governance as well as software. Manufacturers need agreement on which outcomes matter at plant level, who can act when performance moves outside an acceptable range and how local decisions affect enterprise commitments. Without that structure, a unified data platform can become another source of alerts rather than a mechanism for coordinated action.
Culture also determines whether employees see intelligent systems as useful support or a threat. Linke argues that manufacturers should frame AI as a way to augment judgment, preserve scarce expertise and remove low-value work. That becomes particularly important as experienced operators retire and plants struggle to replace knowledge accumulated over decades.
“We are not talking about removing people from the operation,” he says. “We are talking about giving them better tools and making them more capable of taking decisions. When people work with AI capabilities and use them to focus on the most critical issues, the outcome should be better for both the employee and the operation.”
AI must be attached to an operational outcome
The expansion of industrial AI creates a risk that manufacturers repeat the dashboard problem at a more sophisticated level. A model may generate an accurate prediction, but it creates little value unless the organization knows what decision it should trigger, who is responsible and how the intervention will be measured.
Linke groups useful applications around three operational goals: making assets work harder, helping people work smarter and making processes more efficient. The categories overlap, but each begins with an existing manufacturing need rather than a search for somewhere to deploy AI.
Asset-focused applications can help sustain output and reduce unplanned downtime. Workforce applications can place the experience of a long-serving expert in the hands of a less experienced operator. Process applications can automate routine work, coordinate remote support and allow specialists to oversee several facilities.
At Borouge International, a petrochemicals producer operating the Ruwais complex in Abu Dhabi, this progression is already moving toward autonomous operations. Honeywell Technologies has demonstrated an AI-enabled control platform that can detect abnormal situations, recommend or take action and support operators in managing complex process conditions. During multiple pilots, the system made predictions an average of five to ten minutes before alarm incidents, creating time to address the condition before it escalated.
The significance is not the removal of the control-room operator. It is the transfer of repetitive cognitive work to an agent that can monitor more variables continuously than a human team, while operators retain responsibility for higher-value decisions. Connected operations become more resilient when the system recognizes the early stages of a problem and places a response into the workflow rather than adding another alarm.
From connected assets to coordinated operations
Over the next several years, AI is likely to sit more closely alongside production, maintenance and quality processes. The manufacturers that gain most will not necessarily have the largest volume of data, but they will connect information to responsibility and translate insight into a repeatable response.
That will require continued investment in integration, but also discipline around workflows, decision rights and the role of people. Data must be available across functions without stripping away the context that gives it meaning. AI must be deployed where it improves an operational outcome, not because a new tool has become available.
The factory beyond visibility is not one in which every activity is autonomous. It is one in which people, assets and processes operate from a shared understanding of what is happening and what should happen next. When information moves directly into action, connectivity stops being an infrastructure achievement and becomes a manufacturing capability.
