Why Billion-Dollar AI Investments Are Failing to Deliver and How Structured Data Changes Everything
The gold rush is on. From Silicon Valley to the DAX, enterprises are pouring billions into Artificial Intelligence visible in the soaring valuations of chipmakers, the scramble to embed Large Language Models (LLMs) into every software stack, and the relentless pressure on leadership to “do something with AI.”
But as we move deeper into 2026, a sobering pattern is emerging: organisations are investing in powerful AI engines while neglecting the fuel those engines run on. The result? Brilliant technology, underwhelming results.
At VERZE.ai, we have been starting to work on a solution for this problem since 2017, long before “Generative AI” became a boardroom buzzword. Our conviction then, as now, is straightforward: an algorithm is only as good as the data structure beneath it.
The Great Data Starvation
The industry is quietly hitting a data wall. While organisations race to deploy the latest neural networks, the underlying data powering those networks often remains siloed, inconsistent, and poorly structured.
According to Gartner, the absence of “AI-ready data” is now the primary reason AI projects fail to deliver ROI — not a lack of compute power, not a shortage of talent, not insufficient model sophistication.
The Core Paradox
The most sophisticated algorithms from Meta, Google, and OpenAI are “starving.” They are capable of remarkable things, but only when fed high-quality, relevant, structured content. When given poor data instead, the AI does not simply produce lesser results. It hallucinates, misinforms, and generates what one industry analyst calls “fast, wrong, and invisible” errors: mistakes that are confident, plausible, and deeply damaging to brand trust.
The cruel irony is that bad data does not announce itself. An AI running on junk input rarely returns an error message, it returns an answer. That is the danger.
From Chatbots to Agentic Action: The Stakes Are Rising
The AI conversation has moved decisively beyond chat. At NVIDIA’s GTC conference, Jensen Huang emphasised that the next wave of AI is “agentic”: autonomous systems capable of reasoning, planning, and executing complex, multi-step tasks without constant human direction.
“For an AI agent to navigate your procurement process or marketing strategy, it does not need a better prompt. It needs a map.”
That map is built from structured data. Bain & Company has noted that the organisations winning in the AI era are those that have laid a solid data foundation enabling what they term “Commercial Excellence.” Without that foundation, even the most capable AI model is, as one analogy puts it, a Ferrari engine stuck in a mud pit; raw power with nowhere to go.
As AI systems are increasingly trusted to take real-world actions — drafting contracts, managing supplier relationships, triggering procurement workflows — the cost of feeding them bad data is no longer measured in poor recommendations. It is measured in operational failures.
The VERZE.ai Advantage: A Nine-Year Head Start
Most organisations are only now realising they need to rebuild their data governance infrastructure for the AI age. KPMG has observed that this is not a technical upgrade, it is a strategic transformation, one that takes years to execute properly.
We started in 2017.
The VERZE.ai platform was built from the ground up around a “Structured Data First” philosophy, not because we anticipated a generative AI boom, but because we understood a fundamental truth about data: structure is what transforms raw information into usable intelligence. That principle holds whether the consumer of that data is a human analyst or a language model.
What we offer today is not a roadmap or a beta. It is a battle-tested platform already helping organisations capture what McKinsey calls the “Data Dividend”: the tangible, compounding value unlocked when generative AI is powered by high-quality, well-structured enterprise data.
What “Structure First” Means in Practice
It means your data is clean, complete, and consistently formatted before it reaches any Algorithm or AI model. It means your product catalogues, contracts, supplier records, and customer data, marketing messages and more are not just stored, they are organised in a way that machines can reason over reliably. It means your AI initiatives rest on a foundation, not on sand.
Why This Cannot Wait
The cost of inaction is compounding. Gartner projects that by 2030, all IT work will involve AI in some capacity. Organisations that fail to address AI-readiness today will not simply fall behind, they risk structural obsolescence.
KPMG has framed the “Data Quality Trifecta” — completeness, accuracy, and structure — not as an IT concern but as a CEO-level strategic imperative. The organisations that treat data quality as a back-office issue will discover, too late, that it is in fact a competitive moat — or a competitive liability.
You cannot prompt your way out of a bad data strategy. No model update, no fine-tuning exercise, no vendor contract will compensate for data that is unstructured, inconsistent, or incomplete at its source.
The Message to Leadership
Stop focusing solely on the “brain” — the AI model — and start investing in the “blood”: the data that gives it life. Billion-dollar algorithms are only as intelligent as the information they process. The organisations that understand this now will define the competitive landscape for the rest of the decade.
At VERZE.ai, we have the tools, the architecture, and nearly a decade of proven experience to bridge this gap for marketing and communication teams. The world is catching up to what we have known since 2017. The question is whether your data will be ready when it does.
If you are ready to turn your data into a genuine AI advantage, we would welcome the conversation.
References
Bain & Company. (2025a). Building the foundation for agentic AI: Technology report 2025. https://www.bain.com/insights/building-the-foundation-for-agentic-ai-technology-report-2025/
Bain & Company. (2025b). Why AI stumbles without a solid data strategy. https://www.bain.com/insights/why-ai-stumbles-without-a-solid-data-strategy/
Digit. (2026, March). NVIDIA GTC 2026: Jensen Huang says AI future goes beyond just chat. https://www.digit.in/features/general/nvidia-gtc-2026-jensen-huang-ai-future-goes-beyond-just-chat.html
Gartner. (2025a, February 26). Lack of AI-ready data puts AI projects at risk [Press release]. https://www.gartner.com/en/newsroom/press-releases/2025-02-26-lack-of-ai-ready-data-puts-ai-projects-at-risk
Gartner. (2025b, August 5). Gartner Hype Cycle identifies top AI innovations in 2025 [Press release]. https://www.gartner.com/en/newsroom/press-releases/2025-08-05-gartner-hype-cycle-identifies-top-ai-innovations-in-2025
Gartner. (2025c, October 20). Gartner survey finds all IT work will involve AI by 2030 [Press release]. https://www.gartner.com/en/newsroom/press-releases/2025-10-20-gartner-survey-finds-all-it-work-will-involve-ai-by-2030
IBM. (2025). AI data quality: Ensuring your data is ready for the age of AI. https://www.ibm.com/think/topics/ai-data-quality
KPMG. (2025a). Rebuilding data governance in the age of AI. https://kpmg.com/us/en/articles/2025/rebuilding-data-governance-in-age-of-ai.html
KPMG. (2025b). The data quality trifecta: Fueling AI success. https://kpmg.com/in/en/insights/2025/10/data-quality-trifecta.html
McKinsey & Company. (2025). The data dividend: Fueling generative AI. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-data-dividend-fueling-generative-ai
NVIDIA. (2025). NVIDIA AI data platform: Accelerating the data pipeline. https://www.nvidia.com/en-us/data-center/ai-data-platform/
Spend Matters. (2025, August 5). Fast, wrong, and invisible: How bad data destroys good AI. https://spendmatters.com/2025/08/05/fast-wrong-and-invisible-how-bad-data-destroys-good-ai/
Turning Data Into Wisdom. (2025). 70% of AI projects fail, but not for the reason you think. https://www.turningdataintowisdom.com/70-of-ai-projects-fail-but-not-for-the-reason-you-think/