This episode of Nerd @ Work Lab arrives just before Christmas, a little later than planned, after a turbulent period that has finally settled down. That delay is not an excuse, but a context. It mirrors the very theme of the conversation: a shared sense of acceleration, pressure, and cognitive overload that many of us are experiencing in our professional lives.
The discussion with Erick Mahle revolves around artificial intelligence, but not in the superficial sense of tools, demos, or announcements. Instead, it focuses on what it really means to adopt AI inside organizations, especially when optimism runs faster than reality.
AI optimism and the maturity gap
In the US market, optimism around AI is undeniable. Every company wants it, every board asks for it, and every roadmap seems to include it. Yet wanting AI and being ready for AI are two very different things.
Erick introduces the idea of an AI maturity pyramid. At the bottom are basic use cases: copilots, chat assistants, tools that help write emails or summarize content. Almost anyone can start there. At the top, however, are advanced scenarios: autonomous agents, outbound AI, predictive systems deeply integrated into business processes. Reaching those levels requires something far less glamorous than hype: solid data, clean models, and a strong infrastructure.
Many organizations are still stuck at the lower levels. Not because they lack ambition, but because they lack foundations. Data is often missing, fragmented, or unreliable. In some cases, companies want predictive systems without even collecting the data required to build them. The result is frustration, stalled projects, and a growing disconnect between vision and execution.
Boards, pressure, and disconnected expectations
A recurring pattern emerges: boards of directors push for AI adoption because other companies are doing it. Fear of missing out becomes a strategic driver. Meanwhile, the people working on the ground know the limitations of their systems and processes.
This creates tension. AI initiatives are launched without a clear understanding of readiness, leading to rushed investments and unrealistic expectations. The problem is not ambition, but the lack of alignment between strategy and operational reality.
From hype to business cases
One of the strongest themes of the episode is discipline. Every AI initiative should start with a business case. Not just a technical proof of concept, but a clear answer to two questions: does this make commercial sense, and do customers actually want it?
In some scenarios, AI works extremely well. For example, automating simple customer service interactions can reduce waiting times and improve efficiency. In other cases, customers still prefer human interaction, especially when complexity or trust is involved. Forcing AI where it is not welcome only burns money and damages experience.
Success is measured through concrete KPIs: handle time, productivity, cost reduction, and customer satisfaction. Some results appear quickly, others take months. What matters is having a baseline and the patience to evaluate impact over time.
The AI bubble: real, but misunderstood
The conversation openly addresses the idea of an AI bubble. The conclusion is nuanced. AI is not a passing trend. It is here to stay. But current market valuations and consumption forecasts assume that most companies will be able to fully exploit AI in the near future.
That assumption is fragile.
Many organizations are simply not ready. Consumption-based pricing models add uncertainty, making costs unpredictable and procurement more complex. Under pressure, companies sign contracts to avoid falling behind, only to realize later that they cannot sustain or justify the investment.
A correction is likely. Not a rejection of AI, but a rebalancing. Companies that invested early in data and infrastructure will move forward. Others will need to step back, fix fundamentals, and rebuild.
Dreamforce, vision, and reality
Large events and CEO announcements play a specific role. They present a vision of what is possible, not necessarily what is immediately usable. Dreamforce, in this sense, is aspirational by design.
For practical, day-to-day improvements, professionals must look elsewhere: documentation, communities, smaller events, and hands-on experimentation. Waiting for big stages to solve everyday problems is a mistake.
We are all overwhelmed
The closing message is perhaps the most human one. The speed of change is exhausting. New tools appear every week. Skills become obsolete in months. Feeling overwhelmed is not a personal failure. It is the default state of the industry.
Everyone is on the same journey.
The worst reaction is paralysis. The best one is curiosity. Pick one tool. Experiment. Learn the patterns. Once you understand the landscape, switching tools becomes easier. Mastery is temporary, but understanding endures.
AI will keep changing. The only sustainable strategy is to keep moving, one step at a time, hands dirty with ideas, just as the Nerd @ Work Lab was meant to be.
