Latest / The Manufacturing Podcast with Fexingo: Factories, Supply Chains, and Industrial Business / The Factory Floor Automation Paradox
Transcript
- Lucas: Look, if you have been in manufacturing long enough, you know the default assumption is always 'more automation is better.' We spend billions every year on robotics, conveyor systems, and automated guided vehicles because the logic seems sound: machines don't get tired, they don't call in sick, and theoretically, they never miss a beat. Luna: Right, but theory rarely accounts for the messiness of real-world demand shifts, especially when customers want smaller batches delivered faster. Lucas: Exactly. And that is where we see this fascinating, counter-intuitive trend emerging in Q three of twenty twenty six. A lot of mid-sized industrial players are actually de-automating specific stations. They are ripping out expensive robotic cells not because the tech failed, but because it became too rigid to handle the new reality of high-mix, low-volume orders. Luna: So they are trading speed for adaptability? That sounds like a massive capex write-off if the equipment was just bought two years ago. Lucas: It is painful, but let me give you a concrete example from a Tier Two automotive supplier in the Midwest. They had a welding station fully automated with six articulated arms. It was designed for a run rate of five hundred identical units per shift. When market demand shifted toward custom configurations, that same station took forty minutes to reprogram and physically retool between jobs. A human welder could switch types in four minutes. Luna: That is a tenfold difference in changeover time. So the robot is faster at the job, but slower at the context switch? Lucas: Precisely. The bottleneck wasn't the cycle time; it was the setup time. When you are running fifty different part numbers across the week, that forty-minute delay accumulates into hours of lost capacity daily. The plant ended up losing more money on idle downtime during changeovers than they were saving on labor or scrap rates. Luna: I have seen similar issues in electronics assembly where the programming overhead for custom fixtures killed the margin on small batch runs. Lucas: That is the key insight here. Most capital allocation models focus entirely on unit cost reduction. They calculate savings based on volume assumptions that held true in twenty twenty two, but they do not factor in the penalty of rigidity. When you automate for efficiency, you often sacrifice flexibility, which is now the primary competitive advantage. Luna: But surely there is a middle ground? You cannot just put everyone back on manual benches for everything. Lucas: There is a middle ground, and it involves a concept called 'collaborative autonomy.' Instead of full automation, these factories are moving toward hybrid cells. They keep the repetitive, heavy-lifting tasks for cobots—those collaborative robots that work alongside humans—but they leave the fine motor skills, visual inspection, and rapid adjustment phases to the operators. Luna: So the human becomes the software engineer of the physical process rather than just a pair of hands? Lucas: That is a great way to put it. The human handles the variability. The robot handles the fatigue. In that automotive case I mentioned, they replaced the six-arm cell with a single cobot arm that assists with holding heavy brackets while the human performs the final weld. The cycle time went up by twelve percent, but the changeover time dropped to under five minutes. Luna: Which means overall throughput for mixed orders actually increased, even if pure volume output slightly decreased? Lucas: Correct. Because they could process more distinct order types without stopping. The factory floor effectively became a series of micro-factories rather than one monolithic assembly line. This approach requires a different skill set from the workforce, obviously. You need technicians who understand both mechanics and basic coding, not just button-pushers. Luna: And that ties directly into the skilled labor gap we discussed last month. Are companies finding it hard to hire for those hybrid roles? Lucas: They are struggling, but it is easier than trying to hire six separate robotic maintenance engineers. The training burden shifts from specialized vendor support to generalist upskilling. Many of these plants are partnering with local community colleges to create certificate programs specifically for 'hybrid manufacturing operations.' It is a practical, grounded solution to a very abstract problem. Luna: It feels like we are finally moving past the hype cycle where automation was treated as a magic wand for all operational problems. Lucas: We are. The narrative has shifted from 'lights-out manufacturing' to 'agile manufacturing.' Investors are starting to ask harder questions about capex returns on purely automated lines. If a factory cannot pivot its product mix within a week, it is considered a liability in today's supply chain environment, regardless of how cheaply it produces standard items. Luna: Do you think this trend will accelerate in the next quarter, or are we already seeing the peak of these retrofits? Lucas: I think we are just getting started. There is a wave of debt maturities coming up for companies that over-leveraged on automation projects in twenty twenty three. As those loans come due, management teams are realizing they built white elephants. They are being forced to rationalize their asset base, and de-automating inefficient stations is a quick way to improve cash flow metrics. Luna: So financial pressure is driving operational pragmatism. That makes sense. Lucas: It does. And it forces a re-evaluation of what 'efficiency' actually means. If you measure efficiency solely by units per hour, you miss the bigger picture of revenue per square foot across diverse products. The latter metric is what keeps the lights on when consumer tastes shift rapidly. Luna: Are there any sectors where full automation still makes absolute sense, or is this hybrid model winning everywhere? Lucas: Heavy process industries still lean heavily toward full automation. Chemical plants, steel mills, and semiconductor fabs have such extreme consistency requirements and safety hazards that human intervention is either impossible or prohibitively risky. But in discrete manufacturing—the kind that builds cars, appliances, and furniture—flexibility is king. Luna: So the rule of thumb is becoming: automate the constant, humanize the variable? Lucas: That is exactly it. If a task is identical from January to December, let the machine do it. If a task requires judgment, adaptation, or frequent tweaking, keep a human in the loop. It is a simple heuristic, but applying it rigorously has saved several companies from catastrophic downtime events this year. Luna: I wonder how many of these decision-makers are willing to admit they made a mistake by going too far the other way initially. Lucas: Very few publicly, but internally, the conversations are quite candid. They realize that the dream of the unmanned factory was a fantasy that ignored the complexity of global trade. Supply chains are fragile, and rigid factories amplify that fragility instead of absorbing it. Luna: It is interesting to see technology serving the business model again, rather than the business model twisting itself into knots to serve the technology. Lucas: That is the ultimate goal. Technology should be invisible and enabling. When it draws attention to itself by causing delays or requiring exotic maintenance crews, it has failed its primary purpose. Luna: Before we dive back into the data on retrofitting timelines, I want to pause for a second. Honestly, if today was worth a coffee to you, that is the link — buy me a coffee dot com slash fexingo. It helps us keep these deep dives ad-free and focused on the stuff that actually matters to your work. Lucas: Yeah, we appreciate it. It really keeps the research rigorous without needing to chase sponsor dollars for every segment. Let's get back to the numbers. Luna: So, talking about those retrofit timelines, how long does it actually take to go from a fully automated line to a hybrid cell? Lucas: Surprisingly fast. Because you are not tearing down the whole building, you are just swapping end of arm tooling and reprogramming logic controllers. Most successful transitions happen within sixty to ninety days of the initial decision, assuming the company doesn't wait for a major capital budget cycle to approve the change. Luna: Sixty days is barely a blink in industrial planning terms. Why do so many firms drag their feet for years? Lucas: Fear of disruption. Managers worry that taking robots offline will cause immediate production misses. But the data shows that temporary dips in raw volume are quickly offset by gains in order acceptance rates and customer satisfaction. People prefer receiving the right product on time over a perfect product that arrives late. Luna: That is a crucial distinction. Reliability beats theoretical maximum speed every time in B2B relationships. Lucas: Absolutely. And it changes how procurement operates. They stop buying robots based on price per unit and start evaluating vendors based on integration speed and reprogrammability. It is a fundamental shift in purchasing criteria. Luna: Do you see any resistance from the engineering teams who built these automated systems originally? Like, is there an ego issue involved? Lucas: Sometimes. There are engineers who built their careers on large-scale automation projects. Admitting that less is more can feel like a personal professional failure. But smart leadership frames it as evolution, not regression. They highlight the ingenuity required to make hybrid systems work, which is arguably harder than installing a static robot. Luna: I like that framing. It turns a retreat into a tactical maneuver. Lucas: Exactly. It preserves morale while fixing the P&L. And ultimately, that is what we are looking for. Sustainable, adaptable growth rather than brittle, high-speed stagnation. Luna: As we wrap up, what is one thing plant managers should check on their floor tomorrow morning regarding this hybrid approach? Lucas: Ask them to measure the changeover time for their top five most complex products. If those numbers are high, you have found your first candidate for de-automation. Start there, before you touch the high-volume, simple items. Luna: Simple, actionable advice. Check the complex stuff first. Lucas: Right. Because that is where the pain is hiding. And solving it gives you the most immediate relief to your bottom line. Luna: Good point. We will see how many companies make that shift before the end of the fiscal year. Lucas: If they move fast enough, they might just survive the next cycle. If they don't, they will be stuck with robots that no one knows how to program.