The global market and AI industry alike are reverberating in the wake of a survey of 4,454 CEOs, 56 percent of whom said that their AI investments neither reduced expenses or boosted revenues. This follows a major MIT study from Sept. 2025 revealing that 95% of AI integration efforts are, according to The Economic Times, “failing to yield significant revenue acceleration.”
The underperformance of AI may be news to investors, who continue pouring money into leading AI growth stocks, but to those working with the technology, it’s unfortunate but not a major surprise.
Why AI Investments Aren’t Paying Off
Tolstoy’s Anna Karenina begins with the famous quote, “All happy families are alike; every unhappy family is unhappy in its own way.” The same is true of AI integration. All the success stories are similar, every failure is unique.
Within CBIZ’s AI practice, we often hear cautionary tales of businesses that rush to adopt AI, spending fortunes along the way, and ultimately pull the plug or abandon these programs without realizing value. There are many reasons AI integration efforts may fail. They may be poorly devised, overlook essential aspects like training, or simply not apply AI towards those problems it is best equipped to solve.
No one factor universally accounts for the failure of 95% of businesses that do not generate revenue from AI, or the 56% of CEOs that aren’t realizing value from AI investments. Still, it’s easy to identify one thing underlying the success of each company and every CEO that does generate revenue from AI: sophisticated data stewardship.
That means collecting, grooming, and organizing data with the AI’s needs in mind.
Think Like a Machine
The prevalence of thought leadership focusing on a future-first version of AI as an autonomous entity that is able to think, reason, and discern like a person may be doing more harm than good. That’s simply not a reflection of where the technology stands today.
AI in its current form is more like an excel data sheet with superpowers. It can generate a lot more than charts and do more than solve simple equations. But like Excel, AI remains entirely dependent on the information provided to it — in both the prompt, request, or instruction, and the data it analyzes in order to generate its response.
Keep in mind that AI doesn’t speak any human languages, instead it is programmed to identify “tokens” that represent small chunks of text, associate those with “tags” that define their function in the context of a sentence and recognize certain specific proper nouns and date formats. Then, it processes that string of information based on the request and generates a language-based response. In short, AI is like the archetype of a character “born yesterday,” in that it must be fed everything it knows by a user or organization before it can be put to work. Only with carefully curated data can AI accomplish the tasks assigned to it.
The dependency of AI on manicured data is not widely understood. Building your AI implementation effort around the interplay between data and AI is essential. Without that as a broadly accepted first principle, it’s easy to see why so many AI initiatives fail to generate value.
How Data Functions for AI
Imagine a business leader examining inventory data over the past quarter. They may immediately hone in on an anomalous entry recording a value of 10x the prior month and mark that as an obvious mistake in the placement of an integer. But what’s really happening is that this person is leveraging institutional knowledge, years of experience with similar data, and mathematical reasoning to spot the error. Even when it appears to happen in an instant, as if by pure intuition, there is a lifetime of prerequisite knowledge at work.
By comparison, even the most sophisticated AI platform can only rely on the context provided to it. That’s why aggregating, cleaning, and organizing data is so essential for AI success. After all, you wouldn’t build a house on a foundation full of holes.
AI can highlight the same anomaly as our fictional business leader above but will do so only if it has the context necessary to recognize the misplaced integer as a potential problem. In addition, applying AI to this sort of task would require that it have ample past data as a reference and point of comparison, as well as a prompt instructing the AI to flag anomalous entries. That’s one reason AI users are encouraged to give AI parameters like numerical ranges when possible.
Effective Data Practices for AI
Every organization is different, with unique data sources, collection practices, and recording formats that make up their approach to data aggregation and management. Whether your business is refining or adopting better data practices, the starting point is always the same: data education and enablement.
With an understanding of how important data is to your organization, and by incorporating best practices for managing it, you can start building a foundation for AI success.
Next, you’ll need to make sure data is clean, accurate, and uniform. Thankfully, technological tools make this process easier than pouring over and retroactively correcting years of old, potentially error strewn data entries.
Once your data is in order, you can use methodologies borrowing from the principles of data science to validate that data. Once your data is centralized, is validated as accurate, and structured in a way that suits AI, you’re well on your way to joining the 5% of businesses and 44% of CEOs that generate tangible value from their AI investments.
Conclusion
With any technology the most common failures are the result of user error. AI is no different and failing to provide useful data to AI systems can be a particularly costly form of user error. CBIZ offers a suite of data management, analytics, and automation services to transform complex, multi-source information into the foundation for AI success. To learn more, contact us today.
© Copyright CBIZ, Inc. All rights reserved. Use of the material contained herein without the express written consent of the firms is prohibited by law. This publication is distributed with the understanding that CBIZ is not rendering legal, accounting or other professional advice. The reader is advised to contact a tax professional prior to taking any action based upon this information. CBIZ assumes no liability whatsoever in connection with the use of this information and assumes no obligation to inform the reader of any changes in tax laws or other factors that could affect the information contained herein. Material contained in this publication is informational and promotional in nature and not intended to be specific financial, tax or consulting advice. Readers are advised to seek professional consultation regarding circumstances affecting their organization.
“CBIZ” is the brand name under which CBIZ CPAs P.C. and CBIZ, Inc. and its subsidiaries, including CBIZ Advisors, LLC, provide professional services. CBIZ CPAs P.C. and CBIZ, Inc. (and its subsidiaries) practice as an alternative practice structure in accordance with the AICPA Code of Professional Conduct and applicable law, regulations, and professional standards. CBIZ CPAs P.C. is a licensed independent CPA firm that provides attest services to its clients. CBIZ, Inc. and its subsidiary entities provide tax, advisory, and consulting services to their clients. CBIZ, Inc. and its subsidiary entities are not licensed CPA firms and, therefore, cannot provide attest services.
