Companies Are Pouring Billions Into AI. Here’s Why They’re Not Seeing Returns

The AI Conundrum: Why Companies Are Struggling to Deliver Results

Despite the significant investments being made in enterprise AI, many companies are finding it challenging to achieve the desired outcomes. The issue lies not with the technology itself, but with the foundation upon which it is built. Poor data quality, inadequate infrastructure, and a lack of clear strategy are hindering the ability of AI to deliver meaningful results. As a result, companies are facing a growing sense of anxiety and frustration, with some even questioning the true value of AI.

Critics of AI, such as cognitive scientist Gary Marcus and tech columnist Ed Zitron, have been vocal about the limitations and potential risks of current AI systems. Marcus argues that large language models are dishonest, unpredictable, and potentially dangerous, while Zitron has described generative AI as a “financial, ecological, and social time bomb.” These concerns are not unfounded, as many companies are rushing to deploy AI without properly addressing the underlying issues.

The Importance of Data Quality and Infrastructure

Andrew Frawley, CEO of Data Axle, believes that the major problem begins before even a single line of code is written. “The performance gap in enterprise AI isn’t a surprise,” he says. “This is what happens when ambition outpaces readiness. Many companies have invested in AI like it’s a product, not a capability, expecting they could flip a switch to unlock immediate value. But AI doesn’t operate in a vacuum. It’s a high-performance engine, and too many are trying to run it on dirty fuel.” Frawley emphasizes the need for companies to establish a solid data infrastructure, including data ownership, governance, and quality standards, before attempting to deploy AI.

Udo Foerster, CEO of German consultancy Advan Team, echoes Frawley’s sentiments, stating that “too often, there’s no clear strategy. No defined goals. Just vague roadmaps and no system for tracking progress.” He also highlights the importance of adapting processes to accommodate AI, rather than simply bolting it on top of existing systems. Foerster estimates that easily 60 to 70% of AI’s growing pains can be attributed to data infrastructure issues.

The Limitations of Current AI Systems

While many enterprise leaders believe that AI can think and make decisions like humans, the reality is that most AI systems still lack the ability to understand context and nuance. This limitation can lead to confident but incorrect decisions, which can erode trust and damage customer relationships. Frawley notes that deploying AI on fragmented or inaccurate data is an act of self-sabotage, as it will amplify existing flaws and introduce a false sense of confidence in misinformed decisions.

Foerster shares a personal example of how AI can go wrong, even in simple applications. He recounts a experience where he tried to book his preferred hotel room, but the AI-powered call center failed to recognize him as a regular guest. The issue was only resolved when a human intervened and turned off the AI system. This anecdote highlights the risks of relying solely on AI, without proper context and nuance.

Building a Foundation for AI Success

The solution to these challenges is not to abandon AI, but to approach it with a clear strategy, proper infrastructure, and a focus on data quality. Frawley points to clients in the telecom and healthcare sectors who have successfully implemented AI by first creating a single customer view and unifying their data. One client saw a 15% improvement in identifying duplicate records, resulting in reduced wasted ad spend and more meaningful customer communications.

Foerster shares a similar success story from the manufacturing sector, where a German firm was able to achieve a 30% reduction in equipment downtime by standardizing machine data and building a central data warehouse. These examples demonstrate that with the right approach, AI can deliver significant value and improve business outcomes.

Conclusion: Moving Beyond the Promise of AI

In conclusion, AI is not a silver bullet that can fix a broken business. Rather, it will expose the underlying issues and amplify existing flaws. Companies must be honest with themselves about their readiness for AI and take a step back to assess their data quality, infrastructure, and strategy. As Frawley notes, “If I had five minutes with a boardroom, I’d ask one question: Would you bet your reputation, your job, and your company’s future on the accuracy of your data? If the answer is anything but an immediate ‘yes,’ the business is not ready for AI.”

Ultimately, the key to unlocking the true potential of AI lies in building a solid foundation, with a focus on data quality, infrastructure, and strategy. By doing so, companies can harness the power of AI to drive meaningful results and improve business outcomes. As Foerster succinctly puts it, “What specific business problem do you expect AI to solve, and how will you measure success? If you can’t answer that, neither the best algorithm nor the biggest budget will help you.”