We as a society have spent years trying to unlock information. Now the harder problem is knowing which information deserves our attention.
I believe the next great data challenge is not access. Data is everywhere. Search engines can surface it in seconds. AI can summarize it, compare it, and answer questions about it. Remember the needle in a haystack? The needles are no longer hidden.
The real challenge now is discernment. It is finding the right data that matters to a particular business, recognizing when it changes, and presenting it to the right person at precisely the right moment.
Global consulting firm McKinsey reported this year that nearly two-thirds of enterprises have experimented with AI agents, yet fewer than 10 percent have scaled them to deliver tangible value. Eight in 10 companies cite data limitations as a roadblock.
The numbers should make us pause. We are building sophisticated intelligence on top of a very old question. What information should this system actually pay attention to?
...
Zero Returns?
I think about it like hiring a brilliant new employee. You give that person extraordinary capabilities, then hand them a desk and say, "Go." What happens next? They still need context. They need to know what matters, what does not, what deserves escalation, and when you need to hear from them. Give a highly capable employee the wrong information, and they can move very quickly in the wrong direction.
AI is no different.
The technology is powerful. It is also being asked to carry assumptions that no system can magically infer. An AI agent does not automatically know what matters to your business, when a meaningful condition has changed, or whether a particular change deserves your attention.
The cost of getting this wrong is increasingly real. A 2025 report by MIT's NANDA initiative found that 95 percent of organizations in its study were seeing zero return from their generative AI investments.
The figure has been widely summarized as "95% of AI pilots fail," although the underlying research and methodology warrant more careful interpretation. Either way, the warning is unmistakable: access to powerful AI does not guarantee business value.
So what should we be building?
A system that understands relevance dynamically. A system that can personalize the signal, connect to live information, watch for meaningful changes, and trigger an alert when the critical moment arrives.
Timing is inseparable from intelligence.
A piece of information delivered two hours after a decision has been made may be accurate but completely useless. The same information delivered two hours before anyone is ready to act can disappear into the noise. The value sits in the moment of truth, when someone can actually do something with it.
Two Futures
There are two futures here. One is a world in which intelligent systems quietly watch the conditions that matter, surface meaningful changes, and give people room to make better decisions.
The other is a world flooded with alerts, dashboards, and AI-generated answers that gradually become background noise. In the first, intelligence earns trust. In the second, people learn to ignore it.
The good news is that the technology needed for the first future already exists. Stop asking how much more AI we can deploy. Ask harder questions: what information actually matters? How current is it? What constitutes a meaningful change? Who needs to know? And how quickly?
Those questions belong in the boardroom, not at the end of an AI implementation checklist. We have spent enough time making information accessible. The problem currently facing businesses is far more precise: knowing what data matters, knowing when it changes, and making sure the right person sees it in time to act.
That is the business data problem worth solving now.
Deborah A. Taylor is founder and CEO of CalQLogic, a Los Angeles-based technology firm specializing in automation and data analytics.
All opinions are the author’s own.