What if the real advantage in AI lies not in having more data, but in having less?

In this episode of the Don’t Panic, It’s Just Data podcast, host Shubhangi Dua, Podcast Producer and B2B Tech Journalist at EM360Tech, sits down with Herb Blecher, Research Director of Data and Analytics at Enterprise Management Associates (EMA). 

This conversation challenges a common belief in enterprise tech – that gathering everything ensures insight. Blecher, alluding to the modern-day AI craze, cautions the enterprise audience that just because you can access vast amounts of unstructured data doesn’t mean you should.

What is the AI Gold Rush & Why It’s Risky?

Unstructured data now fills the enterprise tech space — voice calls, financial documents, customer chats, images, logs, and emails. “With AI and machine learning, we’ve finally figured out how to access and organise it.”

However, Blecher offers a stark reality check. AI doesn’t just increase insight; it increases error. When machines transition from calculating numbers to interpreting tone, images, and incomplete context, the chances for mistakes rise significantly. A blurry comma in a financial document, a misread abbreviation, a misplaced decimal. In low-stakes situations, this is inconvenient. In finance or healthcare, it can be disastrous.

The danger lies not just in faulty outputs, but in confidently flawed outputs. AI doesn’t hesitate as humans do. It doesn’t say, “This seems off.” It fills in gaps, often convincingly. That confidence, Blecher argues, makes governance essential.

The real issue companies face isn’t a lack of data; it’s a lack of careful thought.

Also Read: AI is Making “As-Code” Inevitable

Why Human-in-the-Loop is Imperative?

Governance over hype is the key takeaway from the conversation. AI generating and using data at the same time creates a new situation. In the past, including financial troubles that Blecher experienced directly, human judgment acted as the final protection. Now, companies risk losing that safeguard in their rush to automate.

Dua puts it simply – humans are leaders; AI is the helper.

The enterprises that succeed with unstructured data aren’t the fastest; they are the most thoughtful. They clearly define their questions first, build feedback loops, monitor continuously, and foster a culture of scepticism.

What are the failures? They often look like ambitious automation without safeguards—from flawed document scanning to high-profile AI rollouts like McDonald's testing automated drive-through ordering, where conversational nuance proved more challenging than anticipated.

Tone, ambiguity, and context remain distinguishing human areas.

What Happens Five Years From Now?

Will AI solve data quality issues? No, it will not. However, Blecher believes that data quality problems are here to stay. “What will change is the range of questions we try to answer. As AI develops, companies won’t stop dealing with edge cases; they’ll broaden the edge.”

The future doesn’t promise easy automation. It promises increased capability, increased capacity, along with increased responsibility.

For CFOs and IT leaders investing in AI-driven data strategies, EMA’s Research Director of Data and Analytics has a final message:

Don’t confuse volume with value.

Don’t replace governance with optimism.

Don’t give up scepticism in a gold rush.

AI’s potential is huge. But more data doesn’t always mean better data. In a world eager to gather everything, restraint could be the most radical strategy of all.

Key Takeaways 

  • More data doesn’t guarantee better insights — clarity of purpose matters more than volume.
  • AI doesn’t just scale intelligence; it scales errors if governance is weak.
  • Unstructured data is powerful, but without context and oversight, it becomes a liability.
  • Human judgment remains essential — especially in high-stakes domains like finance and healthcare.
  • The most successful organisations move deliberately, not impulsively, in the AI gold rush.

About Herb Blecher

Herb leads EMA's Data and Analytics practice. He brings more than two decades of experience building solutions across financial services, data product development, and enterprise analytics.

His perspective is shaped by leading national data initiatives for U.S. mortgage servicers and government agencies, as well as driving product innovation and strategy in fast-moving technology environments. 

Herb's research spans enterprise data and analytics, including data architecture and platform modernisation, analytics and integration, governance, and AI/ML platforms.

#AI #DataAnalytics #TechPodcast #B2BTech #DataQuality #UnstructuredData #AIGoldRush #HumanInTheLoop #AICorporate #HerbBlecher #EMAPartners #CFOs #ITLeaders #DataStrategy #DontPanicItsJustData #EM360Tech #PodcastClips #DataInsights