Samuel Holcman: How do you sort out the so-called jargon from real-world practices that work? Do the members of your organization find some business or technology advice utterly confusing? Welcome to Real Talk with Sam Holsman. In this program, we set the record straight and in terms that business people and technology people can understand. Now, here is your host, Sam Holsman.

Sam Holcman: I began my career in 1972, and we were called data processing. What we were doing in 1972 was processing data. Today, most organizations are called information technology or AI divisions or whatever. And do know what we're doing today? What are we doing today? Processing data. Any questions? This is Real Talk. Real talk with Sam Holtzman. I'm Sam Holtzman. And today, I want to build upon an article by Joe Procipio, sorry if I pronounced the last name wrong, no harm intended, I hope. And the article was titled, The End of AI as Human Replacement? Which explores a fascinating recent move by the Ford Motor Company. His core story is this. Ford just rehired 350 experienced graybeard engineers after discovering artificial intelligence and automated systems failed to deliver the desired quality level they were looking for. In other words, a flagship big company bet heavily, heavily on replacing seasoned human talent. with AI and had to reverse course. This is no knock against Ford Motor Company because there's a lot of organizations that are doing the same thing. Joe's article is not just a news recap. It is a set of lessons for executives and technologists who assumed AI would neatly substitute for people. I'm going to use his four lessons. as a springboard into the broader question we care about in enterprise architecture and business architecture. Where does AI actually fit? And where does experienced human judgment and architecture discipline remain irreplaceable? Lesson one from Joe, technology has hard limits. Joe's first lesson is that all technology All technology has limits. Over the last few years, corporate technology leaders raced to cut headcount and then blame those layoffs on AI, often treating AI as a simple, cost-effective labor replacement. Underneath that story, many layoffs were really correcting bad over-hiring decisions from the cheap money era. But AI became the convenient scapegoat and the shiny promise. Soon this magic will outperform even our best people. So let's clear out the gray beards now. What Ford discovered and what Joe emphasizes is that the imagined chart of AI replacement capacity always was going to go up and to the right and eventually flatten. New technology looks magical and cheap at first. But then you run into very real limits in quality, robustness, and content and context understanding. For architects and leaders, this takeaway is simple. You never remove your experienced human control system before you have proven that new technology can reliably match or exceed your required quality threshold. Lesson number two. New does not automatically beat old. Yep. Joe's second lesson is that new does not always beat old. There's been a long-standing bias in technology leadership to assume defending the status quo is always wasteful and that adopting the latest technology quickly is always wise. But in the recent AI wave, some CEOs went further, openly framing AI-driven layoffs as the preemptive strikes getting ahead of the curve. on job loss by firing experienced people before truly understanding AI's impact. This time, that conventional wisdom really misfired. AI, as Joe points out, is actually a technology that rewards technical experience and threaten those who lack human intuitive decision-making skills. If you fire the very people who know how to measure quality, understand systems and reason through exceptions, you're removing the expert AI systems understanding that feeds AI that it needs in order to be trustworthy in production. In enterprise architecture and business architecture terms, they remove the human governance layer and expect the two layer to self-govern. Number three, these faster horses still need real drivers. Joe's third lesson uses the old Henry Ford faster horses story, but with a little bit of a twist. He argues that what we call AI today is not magic. It's actually a cumulative result of decades of improving processing, engineering, and focused work by our best mathematical and scientific minds. You know what? Computers are still computers. You know what? Chips are still chips. And you know what? Computer code is still computer code. AI is, in his words, a kind of faster horse, an extension of what we've been doing, not a completely new species. That matters because if coding were only about correct syntax, then we would have been flooded with ninja coders years ago. AI sometimes makes it feel as if syntax was the only gap that was skilled and unskilled and the difference between the two. But in reality, the durable value lies in understanding systems, tradeoffs, and long-term consequences. We have seen similar hype arcs before. Bitcoin as new money, NFTs as new Art, mobile phones is purely liberation before they became doom scrolling machines. Good Enterprise Architecture and Business Architecture asks, how will this faster horse behave in our existing stable of processes, data and governance, and who's actually holding the reins? Lesson number four, stop trivializing experienced talent. Joe's fourth lesson is about language and respect. Maybe we should stop casually calling seasoned technologists, graybeards. Many of the people swept aside in the name of AI as a replacement weren't pirates, and they weren't engineers. They weren't all bearded, and oh, by the way, they weren't all men. They were the very people who accumulated experience, and that enabled organizations to make sense of new technology, integrate it safely, and recover when it fails. As Ford and others started dialing those phone numbers and having awkward conversations to bring experienced talent back, they also, unfortunately, discovered something else. They now have to pay inflated salaries to reacquire what they prematurely discarded. Joe hints that this inflation in the cost of senior talent was predictable and that more big companies will face similar I'm sorry calls over the next several months. From an architecture governance perspective, this is a cautionary tale about short-term costs cutting versus long-term capability, resilience and development. Now let's put it all together. AI, data, enterprise architecture, business architecture. So what does this mean for us? Sitting in the intersection of enterprise architecture, business architecture and AI strategy. First, Joe's argument reinforces that AI is not a human replacement. It is a capability augmenter, a capability augmenter that still depends on rich, well architected data and seasoned human oversight. Second, it underscores that the real differentiator in organizations is less than a tool and more about the architecture. how you structure data, define processes, embed governance, so that AI can operate safely inside the system. Third, it reminds leaders that you cannot outsource accountability for quality to a model. Someone still has to be responsible for measuring, interpreting, and correcting. And that brings me back to where I started. I began my career in 1972. and we were called data processing. What we were doing in 1972 was processing data. Today, we are called information technology or the AI division or whatever. And do you know what we are doing today? Processing data. Any questions? Now, please reach out to us at www.eacewe.org. or www.bacoe.org for a no-fee discussion. Please reach out to us. We can assist you in getting back to where you need to be. Thanks for listening.

Samuel Holcman: Thank you for tuning in to Real Talk. Be sure to join your host Sam Holtzman again for another edition of our program. We'll have more Real Topics of discussion then.