• by Kelsey White , Yesterday

Companies have spent years trying to get more value from their analytics. Naturally, many are hoping AI will finally unlock the insights they’ve been missing.

It might. But it also might expose just how much of an organization’s reporting has been held together by tribal knowledge, analyst experience, and a surprising amount of manual interpretation.

Most reporting works, until someone starts asking questions

Many organizations don't look at their analytics and think they're in trouble. The dashboards update, reports land in inboxes every week, and executives generally have the information they need to make decisions. From the outside, the reporting process appears to be doing its job.

A closer look often tells a different story.

Analytics environments rarely become inconsistent overnight. Someone implements a new platform. A marketing team starts tracking an event differently. Sales needs another field, or a report gets rebuilt to answer a specific question. None of those decisions seems significant on its own, but over time they add up. By the time someone notices the inconsistencies, analysts have usually figured out how to work around them.

AI has no way of making those judgment calls. It works from the data it's given. If the underlying measurement is inconsistent, the outputs can be inconsistent too, even when the answers sound polished and confident.

That challenge is already showing up in newer forms of measurement. In a recent report, 73% of B2B organizations said they track AI referral traffic, yet only 34% said they highly trust the metrics they're using. As AI becomes a larger part of the marketing technology stack, confidence in the underlying data is becoming just as important as the technology itself.

Attribution doesn’t get easier with AI

Attribution is another example. Organizations have been debating revenue attribution for years, building increasingly sophisticated models to explain which marketing and sales activities influenced a deal. Those models can be valuable, but they also rely on assumptions that aren’t always obvious to the people reading the reports.

Adding AI doesn’t eliminate those assumptions. If anything, it can make them less visible because the answer arrives instantly, wrapped in language that sounds authoritative. That’s a dangerous combination if the underlying logic hasn’t been challenged in years.

Speed only helps when the foundation is solid

One misconception about AI is that it somehow improves the quality of the data it’s analyzing. It doesn’t. If measurement is disciplined, definitions are consistent, and ownership is clear, automation can absolutely help analysts spend less time cleaning data and more time finding meaningful patterns.

If those things aren’t true, AI simply accelerates the confusion.

That’s why the organizations seeing the greatest value from AI often aren’t the ones with the newest tools. They’re the ones that invested in measurement long before today’s AI wave arrived.

Artificial intelligence isn’t creating weaknesses in analytics. It’s making longstanding ones much harder to overlook. For many leaders, that may end up being one of its most valuable contributions.