By Kevin Keenan, Vice President, Corporate Communications, Reltio
The request lands with a chill. The CEO needs clarity and time is running out. Board members press: After investing millions in AI, why is there no significant progress? Radical promises have been reduced to modest pilot projects, half-constructed initiatives and perhaps worst of all: angry customers and a looming PR crisis.
Here’s the uncomfortable truth: Most enterprise AI initiatives get stuck in IT purgatory because they’re too expensive to abandon and too disappointing to implement. An MIT study, for example, found that despite massive investments, 95% of organizations achieve no measurable ROI on AI due to poor implementation and poor data strategies.
This tension reflects a broader pattern playing out in many companies: bold AI ambition colliding headfirst with shaky data foundations. Organizations are eager to leverage agentic AI for real gains, but most are discovering that the technology can’t get past the broken data it hides. Until this gap is closed, even the most promising AI strategies will struggle to move beyond pilot purgatory.
The immense promise of AI: can it keep its promises?
For most leaders, the potential of agentic AI seems both exciting and intimidating. For example, according to a HBR survey of more than 400 global business leaders, 91% believe agentic AI will transform the future of work, and 83% say its effective adoption will be essential to remaining competitive. Yet only 38% believe their organization is well prepared to do so.
This gap between ambition and preparation defines this moment in AI. As Manish Sood, founder and CEO of Reltio, says:
“Few moments offer as much transformative potential – and as much risk – as the rise of agentic AI. The power of technology to reason and act introduces a new level of autonomy, speed and intelligence to business processes. But this power depends entirely on the quality of the data behind it.
As McKinsey also pointed out, “pull the thread out of these (AI) use cases, and it will lead you back to your data.” In a survey, McKinsey found that 72% Large enterprises have identified data management as one of the main challenges preventing them from developing AI use cases.
Data is the great business technology dichotomy of our time. Data is both the most valuable asset and the lowest quality resource for most businesses. This is also a huge potential liability in the AI age, as LLMs are the most confident liars in the tech world. It takes bad information and broken processes and amplifies them.
AI fails to find the truth buried in the mess of enterprise data
Relti
Data locked in individual applications and systems creates real problems for businesses. Once information is distributed across siled tools, it becomes out of sync. Multiple versions of the same document surface in different places and the idea of a single source of truth disappears. That’s why a simple question like “How many new customers did we sign last quarter?” » can produce four different answers in the areas of sales, marketing, finance and IT. People learn not to rely on numbers. An LLM will not admit this confusion – they will not say, “My sources conflict.” It will simply provide a confident answer, even when the underlying data is wrong.
Smart data and context are the answer. Winning companies are already using it
In the age of AI, not all data is created equal. The companies that win will be those that don’t just collect more data: they operationalize it in a contextualized way. smart data.
Intelligent, context-rich data is reliable, continuously updated information mobilized in real time to guide decision-making by humans and AI. It’s the difference between giving your AI agents a bunch of incomprehensible spreadsheets and a clear 360° view of the critical information your business runs on.
Here’s what sets smart data apart:
Without intelligent data and context, AI becomes an expensive science project. AI cannot fix or hide underlying data problems; this amplifies them.
Enterprise Data Rules Are Changing Rapidly
AI is becoming an uncomfortable boardroom topic for data and IT leaders. The investment is there, the returns are not. It doesn’t have to be this way. Agentic AI can be effective; it just needs the right information and context to power it.
Industry leaders are already setting the tone. Global fast food chains, retailers, pharmaceutical companies, hotel brands, financial institutions, manufacturers and insurers are moving quickly to create reliable, real-time databases. They use them to power fraud detection agents, equip customer service co-pilots with accurate, up-to-date profiles, and replace static dashboards with intelligent workflows that can act autonomously.
The new playbook is here. And those who learn the rules first will shape the market.
Explore the new rules of smart data. See how Industry leaders are unifying trusted data to stay ahead in the AI era.
This article was created by Reltio with Insider Studios.