From Schlitz to AI: When Efficiency Eats Trust

Simon Carver

[warmly] Welcome to the show, everybody. I'm Simon Carver, here with Lachlan Reed and Lara Rowan Croft, and it's really good to be with you.

Lachlan Reed

[warmly] G'day, good to have you along. If you're listening in the car, on a walk, or hiding from your inbox for ten minutes, fair play.

Lara Rowan Croft

[calm] It's good to be here with both of you, and with everyone listening.

Simon Carver

So let me start with the sentence that made me stop and blink twice: a 19th-century beer company from Milwaukee may be one of the clearest warnings we have about AI in 2026.

Lachlan Reed

[deadpan] Yep. Beer. Brewing vats. Brown bottles. And somehow it ends with chatbots and model risk. It's like finding out your granddad's shed manual explains cloud architecture.

Lara Rowan Croft

[matter-of-fact] It sounds odd until you look at the pattern. What happened at Joseph Schlitz wasn't really about beer alone. It was about trust, operating discipline, product integrity, and leadership deciding to optimize the wrong variable.

Simon Carver

And that's the thing that got me. This isn't one of those history stories where we politely say, "Well, that's interesting." This one feels uncomfortably current. Because Schlitz built something people believed in... and then, gradually, leadership treated that product like a throughput problem.

Lachlan Reed

[scoffs] Throughput problem -- that's the phrase, isn't it? Like, instead of asking, "Why do people love this thing?" they asked, "How do we shove more of it out the door faster?" Which, look, works right up until it absolutely does NOT. Same as over-tightening a bolt on an old trail bike. Feels efficient for about five seconds, then the thread strips and you're muttering in the shed.

Lara Rowan Croft

What's actually happening in that moment is a shift in leadership logic. Craft becomes process. Process becomes efficiency. Efficiency stops being an outcome and becomes the goal itself. Once that happens, people start removing the very controls that made the product trustworthy.

Simon Carver

Wait -- "efficiency becomes the goal." That's the line I want to sit with. Because nobody in a boardroom says, "Let's ruin the thing people trust." They say, "Can we increase volume? Can we reduce cost? Can we compress time?" Those sound responsible.

Lara Rowan Croft

[reflective] Exactly. This isn't accidental. The language stays rational while the substance degrades. That's why these failures are often hard to see from the inside. The metrics improve first. The damage shows up later.

Lachlan Reed

And that's where AI starts looking a bit spooky, eh? Because right now heaps of companies are staring at output charts -- more code, more content, more replies, more decisions -- and going, "Beauty, we've cracked it." But they haven't stopped to ask whether the thing getting faster is also getting thinner.

Simon Carver

Thinner is such a good word for it. Not obviously broken. Just... less nourishing. Like a meal that looks huge on the plate and somehow leaves you hungry twenty minutes later.

Lara Rowan Croft

Or a deliverable that looks complete but has no depth, no validation, no judgment behind it. In regulated industries, that's not a style issue. That's risk.

Simon Carver

So that's where we're headed today: Schlitz as a case study in trust and craftsmanship, and then the parallel to AI -- where speed can create the illusion of progress while quietly hollowing out quality.

Lachlan Reed

[curious] And maybe the blunt question underneath all of it is this: when you "optimize" a human system, how do you know which bit is the engine... and which bit is the soul?

Simon Carver

All right, let's start with the rise. Schlitz didn't become a giant by accident. It earned a reputation for consistency and quality, and one of the most memorable stories is after the Great Chicago Fire. They sent beer into Chicago. That wasn't just inventory movement -- that was relationship-building at scale.

Lachlan Reed

[warmly] The Great Chicago Fire bit sticks with me. Because that's not some glossy brand campaign. That's a company showing up when a city's hurting. People remember who helped when things went pear-shaped.

Lara Rowan Croft

And trust compounds. If you step back and look at the pattern, Schlitz built market position through reliability and customer confidence. They also improved the actual experience. The brown bottle innovation mattered because it protected beer from light damage. That is product quality made visible.

Simon Carver

Right -- the brown bottle wasn't a marketing flourish. It solved a real problem. And I think that's important because we throw the word "innovation" around now as if any change counts. But this was innovation in service of quality.

Lachlan Reed

Exactly. Brown bottle means the beer survives the trip better. Simple. That's the kind of improvement even a kangaroo could follow... though I'd still somehow trip over the intro. [chuckles] But then the whole thing turns.

Simon Carver

Yeah. Then leadership starts pushing optimization. Fermentation gets shortened. Ingredients get substituted. Chemical fixes enter the picture. And on paper, it's seductive: lower costs, faster production, more throughput.

Lara Rowan Croft

But the product degraded. Not abstractly -- customers noticed. That's the key. The efficiencies were real in operational terms, but they were extracted from the wrong layer of the system. They came out of the product itself.

Lachlan Reed

[reflective] Shortened fermentation is the one that really lands for me. Because fermentation is basically time doing its job. You can stare at the tank all you want, mate -- it still takes the time it takes. Trying to rush it is like pulling a cake out halfway through and icing the top so no one notices the middle's raw.

Simon Carver

That cake image is perfect. Because from the outside it can look done. And that gets us very close to AI. A lot of AI output has that exact quality: convincingly finished surface, undercooked center.

Lara Rowan Croft

[calm] In enterprise settings, I think of that as green output. It appears complete, but it hasn't matured through review, challenge, and contextual judgment. The defects are often hidden: hallucinations, misread intent, unsupported confidence.

Lachlan Reed

Green output -- yeah, that's a keeper. Like green timber. Looks solid till it twists on you later. And we're seeing it everywhere now: teams smashing out code faster, content faster, customer responses faster... but with less fermentation, less second look, less human sniff test.

Simon Carver

Wait -- "less second look" is the thing. Because AI isn't just increasing output; it's compressing the pause where humans used to catch the weirdness. The half-step where someone said, "Hang on, that sentence sounds right, but is it true?"

Lara Rowan Croft

That's precisely the control point that's being removed. And once review becomes performative instead of substantive, organizations begin to confuse speed with capability.

Lachlan Reed

And customers, same as with Schlitz, don't always complain on day one. They just start trusting you a little less. Then a little less again. Then one day the market notices the taste has changed.

Simon Carver

So if the old Schlitz mistake was swapping craft for efficiency, what's the digital version? To me, it's when companies substitute expertise with generic data and act as if the outputs are equivalent.

Lara Rowan Croft

[matter-of-fact] Yes. Data is not insight. Insight requires context, experience, judgment, and consequence awareness. When a company starts relying on AI-generated responses for customer interaction, strategy work, or decision support without that human layer, it begins diluting its own identity.

Lachlan Reed

That word -- identity. That's bigger than accuracy, hey. Because if every company uses the same generic engine the same generic way, eventually everybody sounds like the same beige bloke in a polo shirt.

Simon Carver

[laughs] The Beige Bot. But seriously, that sameness matters. If your customer service, your strategy memos, even your internal decisions all start coming out with the same polished, flattening tone, you've outsourced not just labor but point of view.

Lara Rowan Croft

And then there's the governance issue. The "flake" problem in the Schlitz story is a useful parallel: a shortcut used to fix one issue creates a larger, visible failure. In AI terms, one flawed output feeds another process, which feeds another model, and now the error is no longer isolated. It's cascading.

Lachlan Reed

[exhales sharply] Cascading is the scary bit. One dodgy answer doesn't stay one dodgy answer. It gets copied into a brief, then into a customer message, then into a decision log, and suddenly the whole shop's building on rotten floorboards.

Simon Carver

Let me try to say that back. So it's not just, "AI made a mistake." It's, "AI made a mistake inside a system that treats AI output as clean input for the next step." And that's how one bad sentence becomes operational reality?

Lara Rowan Croft

[calm] Exactly. That's what we would call cascading risk. Without validation, strong data integrity, and clear accountability, AI amplifies preexisting weaknesses. It does not resolve them.

Lachlan Reed

Which is why layering AI on top of broken systems is such a crook idea. It's like bolting a turbo onto an engine that's already leaking oil. You'll go faster... straight toward the breakdown truck.

Simon Carver

There's also the empathy piece. Schlitz didn't just damage the product; it alienated customers in how it communicated. And I think modern AI systems can do the same thing -- efficient, frictionless, and weirdly cold.

Lara Rowan Croft

Or invasive. Or manipulative. Trust isn't only about factual accuracy. It's about whether the interaction feels respectful, proportionate, and human. If a system is technically correct but makes people feel handled, trust still declines.

Lachlan Reed

[resigned] Handled -- oof. That's the word. Nobody wants to feel processed like a parcel at a depot. People can smell when the human's gone missing, even if the sentence structure is spot on.

Simon Carver

So maybe the final lesson is brutally simple. Speed is not the product. Efficiency is not the brand. Those are tools. They're not the thing people trust.

Lara Rowan Croft

The foundation remains human judgment, quality, and trust. AI should amplify those capabilities, not replace them. The organizations that ignore that usually don't fail all at once. They degrade quietly until the market can feel the difference.

Lachlan Reed

[reflective] And by then, you're not fixing a workflow. You're trying to win back a reputation. That's a much harder ride.

Simon Carver

So maybe that's the question to leave hanging: if your company removed the human craft from its AI-enabled work tomorrow, would your customers recognize the difference immediately... or have they already started to?

Lachlan Reed

Thanks for spending a bit of time with us.

Lara Rowan Croft

We'll see you next time.

Simon Carver

Take care.