Racing Into the AI Future: The Hidden Cost of Superficial Strategies
AI strategy must adapt to rapid evolution, market volatility, and global uncertainty—yet most companies fall short.
From most of my discussions on AI strategy with executives and senior leadership, I have often come out underwhelmed. There has been a curious sense of unbridled enthusiasm combined with a bizarre complacency when it comes to AI development, adoption, and integration into their respective businesses. It’s difficult to blame CEOs and leaders of AI startups for being excessive hype sellers if their environments (investors, clients, the public, and stakeholders/counterparts) are all vying for the hype, even demanding it. We are bound to see the disconnect between promises and reality. Two recent cases in point: questiosn on Klarna’s and Figure AI’s claims.
While markets, universities, and the wider world are replete with hype-driven claims, overstated value, and in some cases even snake-oil methods used to create narratives that support extreme valuations, investments, and bets on AI, there are significant real AI opportunities for businesses across the board that need an evidence-based, nuanced, and rational approach. At the top of these priorities should be a principled (data- and evidence-backed, not hype-based), business-aligned, and well-articulated AI strategy.
The events of this past week have been instructional. The market volatility in response to policy changes has been significant, with a potentially irreversible change in the business environment (even if the policy approach was reversed).
When it comes to AI, a direct question arises—were those finance-world participants who tout AI as the deep differentiator for predicting markets ready for this sudden set of events? The answer is an emphatic no. The same applies to all the businesses that adopted AI for use cases such as sales prediction, revenue projections, supply chains, growth estimates, and so on. Business leaders are scrambling to adjust and recalibrate their strategies in this suddenly changed climate (which, even if previously claimed to be predictable, ultimately was not—at least in terms of its scale and scope).
(Sidenote: If you or anyone you know, through AI, predicted these problems before they occurred, and any of your models can provide evidence of predicting this change, I’d love to know (and you’ll prove your claims by substantiating how your business was prepared and remains unaffected in the current environment). I am certain that I won’t hear back on this.)
Of course, the world couldn’t predict this particular set of events and the corresponding responses. These are the so-called edge cases. We argue that predictive technologies are less effective in handling such edge cases or black-swan events. And that’s precisely the point—AI has limits. AI learns from past experiences and, under the current paradigm, generalizes less and less (despite the impressions we get or the takeaways we choose to work with).
The scale of the models can be misleading in suggesting their generalizability, and it’s natural to be confounded by this contradiction.
It’s indeed difficult to address edge cases, and in the current versions of AI applications, these edges remain wide and thick!
There are so many aspects of GenAI adoption that warrant a much more careful AI strategy being put in place. But this isn’t just a critique of AI—it’s a broader lack of foresight across the board when it comes to AI strategy. As I discuss this more broadly with leaders, what I am learning is that:
Unfortunately, business leaders pushing for AI adoption are either uninformed about the challenges or actively choosing to hear what they want to hear.
AI strategies today are so cursory and superficial that they’re often anchored in wishful thinking rather than a rigorous set of evidence-based hypotheses.
Business models driven by the current evolution of AI capabilities can be flimsy, and we will see significant consolidation and failures. For instance:
The barrier to entry for AI-driven products and services keeps getting lowered, and differentiation continues to diminish when companies rely solely on technology as a differentiator.
Technology is evolving much faster. Combined with new entrants and the ability of incumbents to expand into new areas (see, for instance, Mistral’s recent release of Mistral OCR for document parsing as an example, encroaching into Adobe and others’ territories—foundation model developers can now enter and expand into previously niche domains with significantly less investment), guarantees on customer retention—and consequently, recurring, new or expanded revenue—are becoming increasingly unpredictable.
Strategies such as market penetration and switching costs, once considered reliable mechanisms for continued business, are also growing weaker.
With rapid developments in AI capabilities, business model volatility—including costs, margins, and GTM strategies—is destined to increase in the near-to-mid-term. This, in turn, makes investment decisions highly risky.
Internal technology and infrastructure adoption also pose significant challenges due to the above-mentioned factors, among other considerations (such as opportunity costs).
The “urgency” of having AI is overtaking the need for a well-thought-out approach—not just for AI adoption but for its alignment with broader business strategy and vision. As a result, not just AI initiatives suffer in myriad ways, but the strategy is neither comprehensive nor forward-looking.
Most AI strategies I have discussed or observed are busy solving yesterday’s problems—meaning they are largely focused on catching up rather than preparing for the future, let alone proactively shaping it.
Adding to this challenge are macroeconomic and geopolitical uncertainties (which, while visible, remain difficult to quantify)—yet most companies’ AI strategies seem oblivious to these risks.
Companies, startups, and investment communities need to devote much more attention to AI strategy—not just use cases, modeling aspects, or hype-driven narratives that lack evidentiary data.
Are you really on track to realizing AI’s value while safeguarding businesses? While also managing everything unfolding around us? Better revisit your AI strategy!