For years, manufacturers have been encouraged to automate. Faced with labor constraints, rising operating costs and pressure to increase throughput, the logic is understandable. If a process can be automated, doing so can seem like the natural next step.
That momentum isn’t slowing. Deloitte’s 2025 Smart Manufacturing Survey found that 46% of manufacturing executives ranked process automation as a top-two investment priority for the next two years, while 37% said the same for physical automation.
But that raises a more useful question: should it be?
As automation becomes more accessible, the risk is no longer failing to automate quickly enough. Manufacturers also have to consider whether they are introducing technology in places where it adds meaningful value. A robot that eliminates a bottleneck can transform a production process. One added without considering the surrounding workflow, infrastructure or maintenance requirements can unknowingly move that bottleneck somewhere else.
The result is that the most mature automation strategies aren’t necessarily the ones with the most automated equipment. They are the ones that are deliberate about where automation belongs.
1. Start with the problem, not the technology
One of the easiest mistakes to make in automation planning is starting with a solution.
A new robot, autonomous mobile robot (AMR), automated guided vehicle (AGV) or smart system enters the market, and the conversation quickly becomes “Where can we use this?” A better starting point is asking what problem you’re trying to solve.
Look first at where the operation is experiencing friction. Are employees spending significant time on repetitive material movement? Is a particular process creating a safety concern? Are production volumes exceeding what a manual process can reliably support? Is equipment regularly waiting for materials to arrive? These are problems automation may be well-suited to address.
But it’s important to remember that not every manual process is inherently inefficient. Some tasks are highly variable, require judgment or happen too infrequently to justify the investment, infrastructure and maintenance that automation requires. Automating them may increase technical complexity without producing a meaningful improvement in throughput, safety or cost.
Before choosing a technology, establish what success should look like. That could mean reducing cycle time, improving worker safety, increasing equipment availability, addressing a labor constraint or making it easier to scale production. If the team can’t clearly identify the measurable outcome it expects, the investment deserves another look.
2. Look beyond the individual task
Solving the right problem is only the first step. The next is determining how that change affects the rest of the operation.
Imagine automating material movement between two production areas. On paper, replacing manual transport with an AGV or AMR may reduce labor requirements and create a more consistent process. But what happens upstream and downstream? If materials are not ready when the vehicle arrives, the automated system waits. If the next process cannot accept material quickly enough, inventory begins accumulating. If vehicles regularly leave the workflow to charge, additional units may be required to maintain throughput.
The individual task has been automated, but the overall process may not have improved proportionally.
This is why manufacturers should evaluate automation at the system level. Map how materials, people, equipment, power and data move through the facility. A successful automation project improves that broader flow rather than optimizing one step at the expense of another.
Sometimes, the highest-value investment may not be another automated machine at all. It may be improving the infrastructure or process connecting the equipment already in place.
3. Account for the infrastructure
Every new layer of automation places new demands on the facility around it.
More mobile robots increase charging demand. Connected equipment increases dependence on reliable data communication. Higher speeds and duty cycles can change the demands placed on power-delivery systems. Growing fleets may require additional floor space, traffic management and maintenance capacity.
For example, a facility may calculate that it needs a certain number of mobile robots based on travel distance and production volume. But if charging is treated as a separate consideration, vehicles may spend more time unavailable than expected. The operation may compensate by purchasing additional robots when the better solution is redesigning the charging strategy around natural dwell periods in the workflow.
The same principle applies elsewhere. Adding sophisticated controls does little good if unreliable communication prevents equipment from consistently receiving information. Increasing the speed of automated equipment can accelerate wear or strain components if the existing power-delivery system was designed for a different operating profile.
4. Measure performance and plan for what’s next
The evaluation shouldn’t end once the system goes live.
Operating conditions change. Production volumes rise and fall. New equipment enters the facility. Processes that once created bottlenecks may become less important, while new constraints emerge elsewhere. The automation itself also introduces new demands, from maintenance and software updates to employee training and technical support.
Manufacturers should periodically revisit whether automated systems are still delivering the value they were designed to create. That requires looking beyond whether the equipment is technically functioning and assessing whether throughput has improved, downtime has declined, employees are spending less time on low-value work, maintenance demands have increased or assets are sitting idle for significant periods.
But measuring performance is only part of the equation. Manufacturers also need to consider whether today’s automation can adapt to tomorrow’s operation. Product mix, production volumes, facility layouts and technology needs can all change over the life of an automated system. Systems that can be expanded, reconfigured or integrated with different technologies give facilities more flexibility to respond without repeatedly starting over.
That makes ongoing evaluation as much about planning for what comes next as measuring past performance. The answers may justify further automation, but they may also point toward process changes, infrastructure upgrades or better utilization of assets already in place. In either case, the goal is to ensure the automation strategy continues to evolve with the operation rather than locking it into assumptions made years earlier.
Better automation is intentional automation
The question facing manufacturers is not whether automation is good or bad. Automation has already demonstrated its ability to improve safety, productivity, consistency and scalability across industrial environments. The better question is where it earns its place.
That requires resisting the assumption that every manual process is a problem waiting for a robot. Start with the operational constraint. Understand the complete workflow. Make sure the surrounding infrastructure can support the change, then continue measuring performance and adapting as operational needs change.
The goal isn’t to automate everything that can be automated. It’s to make deliberate investments that leave the operation safer, more reliable and more productive than it was before.