The Vegas Chapel Epiphany

While delivering a workshop in Las Vegas recently, I walked past several wedding chapels, each offering quick, convenient ceremonies. Every few feet, another sign advertised a fast, hassle-free wedding. There were even $40 wedding bouquet vending machines.

Yet, 56% of marriages end in divorce. Let’s break that down further. Assume at least 10% stay together primarily for the kids. Another 10%—if we’re being conservative—aren’t exactly living in wedded bliss. That means roughly 76% of marriages are unhappy, strained, or divorced.

If you knew there was a 76% chance of getting hit by a bowling ball when you stepped outside, you’d wear a helmet. And yet, despite these odds, there is an overwhelming social, familial, and governmental push toward marriage. We are encouraged to take the leap—even if statistics suggest we’ll be trying again in five years.

Like many institutionalized systems, marriage persists not because of its success rate but because of deeply ingrained expectations. The same can be said for how organizations adopt AI. We chase the promise of transformation, pour money into the latest technology, and ignore the evidence that most implementations fail.

All this made me reflect on how the systems we design produce the results we deserve. We design marriage as a cultural default, set up incentives that often lead to failure, and then act shocked when divorce rates hover around 50%. We shouldn’t be.

Photo at Caesars Palace (I hope they make it.)

This got me thinking: if marriage is a system with structural flaws, what about corporate AI investments? Like marriage, AI is being pursued with optimism, significant societal (and market) pressure, and little long-term planning. And just like marriage, AI initiatives fail at alarming rates—some studies suggest that over 80% of AI projects never reach deployment or meaningful ROI.

So, what if we applied the same systemic analysis to AI adoption that we should use in marriage? The results are unsettling—but instructive.

The AI Adoption Model: Measured Commitment vs. Hype-Driven Attraction

In societies where arranged marriages are common, unions are formed with deliberate vetting, family alignment, and practical considerations. While they may lack the fireworks of a love marriage, they often yield stability. In contrast, love marriages are based on personal choice and attraction, often pursued with idealistic optimism but without the structural support needed for long-term success.

Many companies adopt AI the way people rush into love marriages—chasing the hype, ignoring foundational compatibility, and assuming it will “just work out.”

Thoughtful AI adoption aligns with business strategy, infrastructure, and governance before committing. It involves vetting vendors, defining measurable success, and ensuring alignment with existing capabilities.
Hype-driven AI success is rarely about the technology itself—it’s about how well the adoption process is structured. Methodical planning trumps blind optimism.

Key Insight for Leaders: AI success depends on continuously adapting without losing the core vision. Just as successful marriages evolve with effort and compromise, AI initiatives succeed when companies integrate AI as a living, evolving component of their business rather than a one-time deployment.

Chasing AI Trends: The Serial Monogamist CEO

Some individuals cycle through relationships, never fully committing. They love the excitement of something new but lack the patience for the deep work required to build a lasting partnership. The problem isn’t that they don’t find great partners—it’s that they never stick with one long enough to make it truly work.

Some CEOs and CDOs behave the same way with technology. They jump from trend to trend—Big Data, Blockchain, AI, and now Generative AI—without fully integrating any of them. Constantly pivoting from one initiative to the next doesn’t allow them to mature.

Companies invest heavily in big data, blockchain, cloud computing, and AI, but rarely build deep expertise in any area.
AI success requires deep adoption, cultural integration, and ongoing refinement—not just another tech fling.

Key Insight for Leaders: AI isn’t a series of short-term flings—it’s about committing to and building on the data stack you already have and evolving from there. True ROI comes from deep investment over time, not serial experimentation.

Planning for the Breakup: AI and Exit Strategies

In our fairy-tale-forward culture, we don’t like to admit it, but all marriages end in death or divorce. Healthy couples have conversations early on about their expectations should things not work out. Others go further and document their understanding of a prenup. This isn’t an expectation of failure; it’s a safeguard against unpredictability. Smart couples define exit terms before emotions cloud judgment.

Companies often treat AI adoption like a lifelong commitment until something goes wrong. They rarely plan for model degradation, ethical risks, regulatory shifts, or vendor lock-in, which leads to wasted resources and expensive, messy disengagements.

AI models decay over time—what happens when performance declines?
Regulations evolve—what if a once-legal AI solution becomes non-compliant?
Vendors change—what if your AI provider stops supporting the system?

Key Insight for Leaders: Just like prenups, exit strategies should be embedded in AI projects. If you don’t plan for failure, failure will plan for you.

AI as an Attention Trap: The Tinder Effect

Modern dating apps create paradox-of-choice exhaustion—so many options that people struggle to commit. People swipe endlessly, always believing something better is just one match away.

Companies treat AI like Tinder. They never commit, constantly testing but never deploying at scale.

They conduct endless proofs of concept but never operationalize them.
They switch vendors frequently, believing the next model will be significantly better.
They expect AI to be plug-and-play, but AI requires training, customization, and patience.

Key Insight for Leaders: AI success isn’t about finding the perfect model—it’s about choosing a direction and iterating over time.

The Midlife Crisis AI Pivot: Overcompensating for the Past

For aging organizations, AI investments are a kind of “Corporate Ozempic”—a quick fix for organizations that ignored foundational fitness for too long (Galloway, 2024). Just as individuals in midlife crises make drastic, impulsive choices to reclaim lost youth, companies that have lagged in AI are now rushing to overcompensate with reckless spending.

Rushing into massive AI investments ($300B) without a roadmap.
Trying to reinvent themselves overnight with AI.
Betting on AI to fix deeply embedded business problems instead of solving root causes.

Key Insight for Leaders: AI should be a measured, continuous investment—not a desperate play to remain relevant. Companies that treat AI as a “miracle cure” rather than an evolving capability risk making costly, ineffective investments.

The Long-Term Cost of AI: The Child Support Effect

Divorces are expensive not just in the short term, but in the ongoing payments—child support, alimony, legal fees, etc.

Companies fail to plan for the true cost of maintaining AI:

AI models degrade and require ongoing retraining (Stanford AI Index, 2023).
AI needs constant compliance monitoring.
AI demands continuous infrastructure costs (Forrester, 2022).

Key Insight for Leaders: AI is not a one-time investment—it’s a perpetual operational cost. Companies must budget for AI’s entire lifecycle, not just its initial deployment.

AI as an Open Marriage: Strategic Flexibility

Some couples thrive in open marriages, balancing commitment with exploration. Similarly, companies must commit to a core AI strategy while remaining flexible enough to test new models and vendors.

They don’t lock into one vendor forever.
They build interoperability into their systems.
They balance stability with innovation.

Key Insight for Leaders: AI strategy doesn’t have to be monogamous—but it must be intentional and structured. The strongest AI programs allow for evolution without losing direction.

Conclusion: Systems Produce Results by Design

The systems we build shape our outcomes. Our approach to marriage has set the stage for high divorce rates, just as our corporate approach to AI has set the stage for high failure rates. We shouldn’t be surprised when hasty, under-planned, and hype-driven AI initiatives fail.

But failure is not inevitable—it is a choice embedded in the way leaders structure their AI strategies. If organizations want different outcomes, they need to design for success, not just experiment and hope for the best.

This requires a shift in leadership mindset:

AI is not a one-time procurement decision—it’s a living investment that must be continuously refined.
AI governance must be as disciplined as financial governance—clear accountability, performance tracking, and adaptability are essential.
AI should be aligned with long-term business objectives, not short-term FOMO-driven adoption cycles.

Leaders who take an intentional, structured, and adaptable approach to AI will build a sustainable advantage. Those who continue to treat AI as a high-stakes gamble—rushing in without clear integration plans—will pay the price.

The future of AI in your organization isn’t a question of capability. It’s a question of leadership discipline.