For years, companies have known their data is valuable—but they’ve only scratched the surface of its potential. With generative AI now entering the mainstream, the urgency to tap into the other 99% of enterprise data has never been higher.
Henrique Lemes, Americas Data Platform Leader at IBM, explains that most businesses still rely on a narrow slice of their available data—primarily structured data that's easy to organize and analyze. But that’s just a fraction of the full picture.
“Less than 1% of enterprise data is currently used by generative AI,” Lemes says, “and over 90% of that data is unstructured.”
This includes emails, documents, videos, social media posts, and audio—formats that are harder to process but rich with insight.
Unlocking this massive volume of unstructured data is essential to improving AI performance, boosting ROI, and enabling smarter decisions. But it’s not as simple as plugging in a new tool. It takes infrastructure, governance, and strategy.
The Data Divide
Structured data is clean, orderly, and software-friendly. It sits in spreadsheets and databases and can be quickly searched or modeled. Unstructured data, by contrast, is messy and diverse. It requires more advanced tools to analyze—and carries more risk if not properly governed.
This complexity has led many organizations to leave most unstructured data untouched. And when data is ignored, trust suffers. Leaders can’t rely on decisions driven by incomplete or unverified information.
That’s where AI-powered data transformation comes in.
Turning the Trickle into a Firehose
Lemes outlines a three-step approach to unlocking data at scale:
- Automated Ingestion – Systems must pull in massive amounts of data, both structured and unstructured, without manual bottlenecks.
- Curation and Governance – Data must be cleaned, categorized, and governed to ensure it’s accurate, secure, and compliant.
- AI Enablement – Once ready, the data can be leveraged by generative AI models to drive better business outcomes.
This process, he says, consistently delivers over 40% ROI when compared to more traditional retrieval-augmented generation (RAG) approaches.
A Smarter Way to Scale AI
IBM helps clients navigate the complexity by aligning tools, people, and processes around a unified data strategy. The goal is to simplify what’s inherently difficult—especially when data is spread across formats, departments, and platforms.
As businesses scale, their data environments grow more complex. Legacy systems often struggle to adapt, especially when AI models expand beyond initial narrow use cases. That’s when unstructured data becomes a make-or-break factor.
IBM addresses this by crafting customized AI roadmaps, prioritizing transparency and compliance alongside capability. From international banks to Fortune 500s, IBM supports enterprises in even the most tightly regulated sectors.
“We prioritize accuracy, governance, and observability,” Lemes says. “That’s how we help clients unlock the full value of their data—at scale and with confidence.”
Sophie Jin
Final Thoughts
The message is clear: the future of AI doesn’t just depend on algorithms—it depends on data. Companies that want to stay competitive need to shift from managing data as a cost center to treating it as a strategic asset. That means investing in the tools, processes, and partnerships that can transform raw information into trusted, actionable insight.
And the good news? The other 99% of your data is finally within reach.
