By Todd Gilliam, North America Industry Leader, CPG, Rockwell Automation
Key takeaways:
Retiring experts take decades of judgment with them. Capture what’s standardizable now, before it’s gone.
Operators need context, not more alarms. Connect production, maintenance, and quality data so the system points to root cause, not just symptoms.
Start with a specific workforce constraint and bring frontline employees in before rollout; 45% of failed AI initiatives cite their exclusion).
Dairy processors face a difficult equation. Demand is growing, plants are adding capacity and operating complexity continues to increase, yet the experienced people needed to run those plants are becoming harder to find and retain. Today, amid ongoing labor challenges, six in 10 U.S. dairy executives say talent is now a top strategic priority.
As a result, many processors are beginning to ask a different question. How can they automate not only production processes but also more of the decision-making processes so a smaller, less experienced workforce can run operations safely, consistently, and confidently?
Done well, this operator-centric automation approach does not replace operator judgment. It gives people better information, clearer context, and easier-to-use tools so they can make confident decisions in increasingly complex environments.
Preserve expertise before it walks out the door
One of the first steps dairy processors need to take is to capture the knowledge of their experienced production personnel while they’re still there. These employees often have unique expertise that is difficult to replace, such as understanding of the subtle variations in raw ingredients and production conditions that impact quality.
Consider a cheese plant preparing for the retirement of an experienced cheesemaker. Over decades, that person has learned how to adjust the process for changes in milk composition, enzymes, suppliers, and seasonal conditions. Some of that knowledge may be documented, but much of it exists as judgment developed through repetition. A new operator cannot absorb 30 or 40 years of experience during a short onboarding period.
The first priority is to separate what can be standardized from what still requires expert judgment. Repeatable tasks can be captured through short videos, visual instructions, approved troubleshooting guides, and step-by-step workflows available directly at the point of work. This can make proven methods easier to access, reduce avoidable variation, and establish a clear escalation path when the situation falls outside normal conditions.
Give operators more context, not more alarms
Many plants already collect large amounts of data, but much of it exists in silos rather than being connected and contextualized to tell the larger story of what’s happening in production. This can create issues like Isolated alarms and dashboards that increase the burden on operators.
For example, if a filler repeatedly stops, the first conclusion is that the filler is the problem. The actual cause could be an upstream jam, a product that does not run well at a certain setting, a quality deviation or a performance issue of a particular shift.
An operator-centered approach connects production, maintenance, quality, and product information so workers can understand the relationships between those systems and identify the likely cause of a problem, not just its symptoms. Instead of asking an operator to search through several systems, call multiple departments, and troubleshoot by trial and error, the system can present the relevant conditions and the next best action.
Simplify the work itself
One way to unburden employees is to design unnecessary complexity out of their jobs.
Standard interfaces, common data structures, and consistent equipment principles reduce the number of systems employees must learn. Instructions placed at the machine shorten the distance between a problem and the information needed to solve it. Modular equipment and preconfigured recipes can reduce the manual work involved in changeovers.
Better automation should also preserve operational knowledge, not just automate repetitive tasks. Model predictive control (MPC) is one example. By capturing how experienced operators respond to changing process conditions and embedding that knowledge into predictive models, plants can reduce unnecessary manual adjustments while helping newer operators make more consistent decisions. These systems can anticipate process variability and recommend the next best course of action, helping maintain product quality while reducing the burden on a less experienced workforce.
Start with a workforce constraint
No single automation project solves a labor shortage. Workforce constraints differ from plant to plant, and different challenges require different solutions.
Plants should begin by defining the specific constraint they want to solve. Is maintenance capacity limiting uptime? Are employees losing time because information is scattered across systems? Is critical process knowledge at risk of disappearing through retirement or turnover?
Once the problem is clear, the plant can select a technology and the appropriate use case for solving it. The project should then be put against an operational outcome such as troubleshooting time, changeover duration, unplanned downtime or quality variation.
Frontline employees should be involved before the solution is selected, not only after it is installed. Research has found that 45% of manufacturing leaders cited the exclusion of frontline leaders from design and rollout as a significant contributor to unsuccessful AI initiatives. The lesson applies more broadly to automation: people who understand the work should help shape how the work is redesigned.
Move people toward higher-value decisions
The dairy plant of the future will ask people to spend less time watching normal production and more time managing exceptions, protecting quality and improving performance.
For dairy processors, that is the real promise of automation in a constrained labor market. The objective is not simply to run with fewer people. It is to preserve the knowledge the plant cannot afford to lose, give every operator clearer information and remove unnecessary complexity from the work. Plants that follow that approach will be better positioned to grow without asking an already stretched workforce to carry more of the burden.
Todd Gilliam currently serves as North America Industry Leader, Consumer Packaged Goods (CPG) at Rockwell Automation. In this role, he leads initiatives to support manufacturing customers in the CPG sector, focusing on automation, digital transformation, and operational efficiency.
