Amy Michtich is the perfect person to talk about the State of Production Health 2026, having paid her dues with more than 30 years in food and beverage manufacturing. She was the first woman to run MillerCoors’ Shenandoah Brewery, rose to Chief Operations and Supply Chain Officer for Molson Coors Canada, led multi-site operations at Scotts Miracle-Gro, and served as a COO in food manufacturing. Recently retired and now consulting, she’s spent her career on plant floors and in the rooms above them, so she’s in a rare position to tell you how the view changes depending on the chair.
It’s also personal: her family came up through coal mining, auto, and steel, all of which have since been decimated, giving her reading of the data a sharper edge when it comes to Augury’s State of Production Health 2026.
This fourth annual report surveyed 501 manufacturing leaders across the US, Germany, France, and the UK. The key finding: the pressure has shifted inside the plant. In 2025, the threats were external, including supply chain volatility and disconnected systems. This year, the top concerns are workforce constraints, now the single biggest limiting factor, and unplanned downtime, which has risen ten points to second place. Meanwhile, AI has moved beyond the pilot stage, with predictive maintenance the most-deployed use case at 57%. However, poor data quality has become the top roadblock, cited by nearly half.
So what does this really mean? We asked Amy.
How do you describe your job to someone who doesn’t have a manufacturing or tech bone in their body?
I spent ten years in brewing, where there were very few women, so I’d just say I was ‘The Beer Lady’ who ran a brewery. They’d go, “Oh, okay,” and that was that. Then I went to a garden company, and “The Soil Lady” didn’t sound as successful. These days I can say supply chain, which people recognize post-COVID, but then I still have to cut to the chase: “We make the stuff that shows up on your shelf at the grocery store.” I’ve had end-to-end procurement, physically through to the shelf. But ‘The Beer Lady’ was always the easier conversation.
Was there anything you wish everyone knew that would have made your job easier?
Never lose sight of your customers. In a plant, you get insular: focused on your KPIs, not the people you’re serving. Leaders who get that do better at moving up in corporate because they see the end-to-end picture. But your people sit in manufacturing, so you think you’re the king, or queen, in my case. Once you see the other levers, you operate differently.
The report’s thesis is that the pressure has shifted from outside the plant to inside it, with workforce and downtime on top. Is this a real shift, or just what leaders are finally willing to name?
It’s a real shift, and it’s ongoing. Consider the four workforce buckets: reskilling, labor shortage, knowledge transfer, and aging workforce. They’ll all continue to worsen. In North America, the technical trades haven’t just aged; they’ve aged out. The craftspeople who were 70 and still working ten years ago are gone, and the institutional knowledge often went with them. After COVID, many plant workers left for other industries, and those who replaced them often came from the service sector, with no intuitive feel for the work.
And the pipeline behind that isn’t reassuring. I’ve got a 19-year-old with zero interest in helping remodel the house. My husband’s a retired industrial contractor, so for him this is barely work. He looks at our kid and says, “Well, good thing he wants to go to college, because he’d never make it on a job site.” In Food, especially, we’re now leaning on AI just to write work instructions, and not in a single language. I might need the same instruction in six languages in a single facility. That’s where a lot of the time goes.
There’s a data-quality paradox: the most mature predictive-maintenance industries, cement and building materials, complain loudest about poor data. Is this what stops you, or a sign you’ve gone far enough to see it? And for a plant manager who hears 47% name it as the roadblock, they may think they have to fix the data first. Or is that a trap?
Honestly, I’m pretty neutral on it. At every place I’ve worked, someone’s told me the data’s inaccurate, or that if you slice it differently, it says something different. That was true no matter what tools we had. We always have more data than we actually use, and we always argue with it before we use it. So for me, it’s just a recurring theme. The question I’d actually put to cement and building materials is whether poor data quality is news to them: Is it a surprise that they’ve maybe been collecting the wrong data for the problems they’re trying to solve now?
For a plant manager, fixing the data first is the trap that runs forever. Most places I’ve worked have had automated data collection for years, integrated into a single dashboard. So, the real question is: Is it broken again? Fix it, then move on. Announcing, “We have poor data quality,” and stopping there is how you stay stuck in analysis.
The C-suite rates its AI progress higher than the directors and VPs running it. You’ve sat in both chairs… Do we need to close the gap?
The people doing the implementation know exactly where they haven’t reached what they need. Meanwhile, the C-suite just sees the project as on track and the dashboard as green. But the executive is also looking wider. A chief supply chain officer may run eleven functions, and much of the live AI work isn’t in manufacturing at all; it’s in planning, demand, and supply, the whole end-to-end. So some of that gap is just where you’re sitting. And I don’t think you need to close it. The people closest to the work should be more critical than those above them. I actually value the gap.
On your home turf… CPG has the lowest predictive-maintenance adoption but the highest supply-chain optimization. Smart sequencing or a blind spot?
Smart and natural. In consumer goods, manufacturing accounts for only about 20 to 25% of the end-to-end budget. Procurement is closer to 60%, and moving finished goods costs more than making them. With stock prices down and tariffs everywhere, you go after the network first because that’s where the real money sits. And because your plants are brick and mortar, they’re much harder to move than a vendor. Most CPG companies have run operational-excellence platforms for years, with predictive maintenance built in as a pillar, so it isn’t being ignored. It’s mature. But meanwhile, the optimization side keeps shifting on you: vendors popping up or closing, the impacts of globalization, and regulations changing what you’re even allowed to buy.
If you had to bet on the one move that separates the manufacturers who win over the next eighteen months from those who don’t?
You don’t get to pick one; you do both: supporting the people on the floor and obsessing over keeping the plants running. Engaging and upskilling your people has always been the job, but the pace now has to be faster than ever because skill loss is huge, the workforce keeps aging, and those coming up behind us aren’t strong communicators yet. On uptime, the move is to put the tools for analyzing downtime into the hands of the one or two people on the floor who can actually run that analysis. Senior leaders should be obsessed with keeping the plants running; the floor should stay focused on getting better at the work. You don’t choose between them.
Last one. If you had everyone’s attention, what would you say?
We’ve thrown away the value of manufacturing in the West, and we need it back. My generation might be the last in North America to have worked in traditional manufacturing and experienced the communities built around it before it all got offshored.
Reshoring has to happen, but what comes back won’t be what left. It’ll be more inventive, and we’re not teaching any of it. We’ve also failed to make the work visible. I do high-school career fairs, and the manufacturing table is empty every time. Kids go to logistics, procurement, and engineering because nobody’s told them the work still happens inside a plant. And I’m saying this, having a 19-year-old who I ask, “What do you want to do?” And then remind him that gaming is not an option.
I won’t pretend it’s glamorous. But it can be a real fit for those who like process and people, combining the strategic with the tactical, and having a different problem to solve every single day. If you want to help run a circus, with a couple of big tents going at once, manufacturing is the place.
People always ask what you’d do if you won the lottery. My answer is easy: establish more trade schools. We stopped building that pathway, and nobody’s rebuilt it yet.
