Transcript of interview with Ranjit Maharajan, head of the Product Group Automation Solutions, ANDRITZ
Jackie Roembke, editor-in-chief, WATT Feed Brands: Hi, everyone. Welcome to the Feed Strategy Podcast. I'm your host, Jackie Roembke, editor-in-chief of Watt's Feed Group. This edition of the Feed Strategy Podcast is brought to you by FeedStrategy.com — your source for the latest news and leading-edge analysis of the global animal feed industry.
Today's guest is Ranjit Maharajan, head of the Product Group Automation Solutions at ANDRITZ. He recently spoke at the Feed Mill of the Future Conference, held at IPPE in January. The sold-out event was organized by Feed Strategy and Feed & Grain magazines in partnership with the American Feed Industry Association.
Today, Ranjit will discuss what he covered during that presentation, with a focus on how feed mills are advancing operator training and decision-making through simulation technology and AI-powered data tools.
Hi, Ranjit. How are you today?
Ranjit Maharajan, head of the Product Group Automation Solutions, ANDRITZ: Hey, Jackie. I'm doing fantastic. How are you?
Roembke: Great. Thank you so much for being here — it's nice to talk with you again. Let's get right into it. Feed mill operators are increasingly working with complex automation systems. What challenges does that create for training, and how are operator training simulators being used to address safety and skill development?
Maharajan: As feed mills become more and more automated, the role of operators has changed considerably. It's no longer just starting and stopping a machine. They have to manage a lot of interconnected systems, understand process behavior, and respond to situations that fall outside of normal conditions. There are always going to be some upset conditions thrown at you.
Today's feed mill operators have to figure out how quickly they can solve these problems. The challenge this creates for training is that many of these situations are rare but critical. Think about it like a pilot — very rarely will he need to land on water, but that training is necessary. You feel comfortable knowing the pilot has gone through it, even though the chances are very slim.
Similarly, there are a lot of things that happen in a feed mill — raw material variability, process upsets, equipment failures — that don't occur every single day. But when they do, the operator needs to be prepared to understand the problem and act on it quickly and safely. That's almost impossible to replicate in a live environment. Today, most training happens through job shadowing or trial and error, and there is real risk to safety, quality, and throughput when you operate that way.
That's where simulation-based training comes in. It gives the operator a realistic environment — much like a flight simulator for pilots. Operators can sit at their operator station and not even know they're running a digital plant. This gives them practice responding to different scenarios, and we can program those scenarios around typical challenges specific to their mill.
What this does is build confidence over time without risking production or safety. We've seen operators repeat multiple scenarios, learn from their mistakes, and be very well prepared when those situations occur in real time. Over time, this leads to more consistent decision-making and a much stronger safety culture. It's a shift from on-the-job training and trial and error to a simulation-based, controlled environment that is structured, repeatable, and safe — which is exactly what modern automated mills need today.
Roembke: During your presentation at the Feed Mill of the Future Conference, you mentioned that operators can now chat with their data using AI tools. What does that look like in practice, and how does it change the way feed mill managers can make operational decisions?
Maharajan: Absolutely. With the advancements in AI and large language models, there are a lot of changes happening very quickly that fundamentally change the way we interact with data at the mill.
In practice, chatting with data is really about changing how people interact with information. Traditionally, a feed mill has a set of alarms, a production report, and a daily report. People tend to react to those at the end of the day — they're being reactive. What AI-driven chat with data does is remove that friction. Operators and managers can simply ask questions in plain language. Everyone is familiar with ChatGPT and Copilot — it's essentially the same technology applied to your feed mill.
For example, a mill manager might say: "I'm noticing throughput is slightly lower than normal, but there are no active alarms — what's going on?" Instead of having an operator or engineer dig through data and build reports to find the root cause, the system does it automatically. It looks at all the historical data, analyzes the patterns, and digs through tons of information from your mill.
It might respond: "Your throughput has decreased 4% over the last two shifts, and that correlates with higher conditioner temperature variability upstream — which is also affecting energy consumption." It might even go further and identify a similar pattern from the past associated with die wear or changes in raw material moisture. That level of insight is available right at your fingertips, in a conversational way.
Instead of reacting late or guessing, the system gives teams the confidence to know exactly what the root cause is — and act on it early, whether that means scheduling a maintenance inspection before quality is impacted or something else entirely. This fundamentally changes decision-making at the mill. Teams can move much more quickly and the data isn't siloed to one or two people.
What I find most valuable is the accessibility. The same data is available to operators, mill managers, and maintenance managers — everyone is working from the same information. It also levels the playing field. An experienced operator may recognize a pattern based on years on the job, but a newer operator would be lost. Now, that newer operator has access to the same data and can make a much more informed decision.
What I've seen over time is that this changes the culture. Teams start asking better questions. When I started using ChatGPT, my prompts got better and better — you learn how to navigate it. It's the same dynamic. You get better at asking questions, the AI becomes a stronger decision-making partner, and it's like having your best operator — with all the data insights — sitting right next to you. That's what chat with data brings to the feed mill.
Roembke: Just to clarify — this is something mills are doing right now, not something down the road?
Maharajan: Correct. It is available now. This technology has been proven in other process industries for quite some time. We already have it available. If someone wants to explore it or experiment, we can absolutely help them.
Roembke: Based on your experience, what are the most common misconceptions about AI, and what results have you seen when companies do take the leap and trust the technology?
Maharajan: This is an excellent question because it comes up constantly. The word "AI" gets applied to everything, and customers are often uncertain what to make of it. The biggest misconception is that they think the system is either too complex or, on the other end of the spectrum, that it's just dashboarding. There's a lot of confusion — and even fear. I remember commissioning a system when an operator asked me point-blank: "Will I still have a job after this is in place?" That fear is real.
Many teams worry about what AI means for experienced operators. In reality, it's quite the opposite. AI enhances operator knowledge and helps them make better decisions. When applied to real operational problems, AI amplifies human expertise — it doesn't replace it. And it produces tangible results.
One example: we installed vibration sensors at a mill. In the feed industry, we don't always have the luxury of analyzing all vibration patterns or access to dedicated vibration experts, so we wait for alarms — which is already too late. When we implemented our system, the AI detected a pattern and flagged early signs of bearing looseness, even though readings were still well within limits. It alerted the operator, the team finished the production run, slowed down, performed a maintenance check, and was able to save two days of downtime that would have resulted from a failure.
Another example involves production throughput and quality. There was a subtle deviation that was almost impossible to track manually. The AI looked not only at current production but at all other parameters — minimum sustainable rate, overall equipment efficiency — and identified a bottleneck in the conditioning stage. Operators proactively adjusted steam levels and were able to optimize the run rather than have that batch become scrap.
What I've seen is that the most significant impact is on decision quality and team confidence. Once teams see that AI can help solve real problems and answer real questions, it becomes a partner rather than a threat. The skepticism drops and adoption grows. The focus shifts from fear and doubt to using the tool to make faster, safer, and more consistent decisions over time.
Roembke: Those are great points, and I think highlighting those practical applications within production is extremely important right now — it can be difficult for operators to imagine without that context. Are there challenges to implementation for feed mills that are interested in exploring this kind of integration?
Maharajan: Yes, and I always say this is a journey. There are many solutions available. The way we approach it is by listening to customers — understanding their challenges and their primary pain points. Based on that and our feed industry expertise, we work with them to put together the right solution. This isn't a product you install and walk away from. You have to collaborate with the mill, understand what they're dealing with, and bring the right technology to the right problem.
One of the real risks is implementing the wrong solution. If it doesn't solve the actual pain point, it becomes an expensive failure and everyone loses confidence. We have to be extremely deliberate about identifying the challenge, confirming we have the technology to solve it, and making sure we're building something that actually works.
And technology is moving quickly. Even if someone looked at these solutions a year ago and didn't see a fit, it may be worth revisiting. The capabilities have grown substantially.
Roembke: When you reflect on the feed mills that have embraced this technology, what is the most surprising transformation you've observed?
Maharajan: What I've personally seen is that it typically starts with low confidence — even some skepticism. And then once teams see the results, there's a moment where they understand: "This is actually helping me." That transformation is what's really striking.
Consider the scale of the challenge. There are roughly 28,000 feed mills around the world — 5,000 to 6,000 in North America alone. If you average six or seven operators per mill and factor in an estimated 15% attrition rate, you're talking about close to 29,000 people churning every year. You simply cannot address that through on-the-job training, job shadowing, or trial and error alone. Add to that the difficulty of recruiting operators in certain locations, and it becomes clear why simulation-based training is so important. It allows people to practice in a virtual environment safely, and the AI system doesn't just train you — it makes you a better operator.
What I've seen over time is that operators gain complete confidence in the technology. People who were strong skeptics early on become its biggest advocates — and they start wanting to expand it to other areas of the plant. In one word, what's remarkable is the cultural transformation. Operators collaborate more effectively, share insights, and trust both their training and the recommendations the AI provides. Skills that used to take years to develop can now be built in months. It elevates the entire team's performance, resulting in a smarter, safer operation.
Roembke: Thank you so much for your time and insights today, Ranjit, and thanks to everyone who tuned in.
