There’s an ad campaign from Google I’ve seen quite a bit lately, mostly via Captivate’s in-elevator video package, promoting Google Cloud. The ad is tied to March Madness, showing basketball players on the court and asking if you know certain stats about the team and its players. It’s not aimed at fans but is geared toward decision-makers and execs.

The ad bugs me as both a marketer and a copywriter. That’s because here’s the tagline for the campaign:

Know what your data knows.

I understand what the sentiment behind the copy is: That you should be able to access the data you have in a way that allows you to make informed decisions. If you are collecting a wide array of pertinent metrics you should be able to dig into those numbers and gain insights that ultimately benefit your business.

Numbers are, with very few exceptions, not knowledge though. That alchemy requires a kind of analysis and comprehension of what the numbers mean and represent.

A simple example of what I’m talking about is just money. If I type “$23.42” the number itself doesn’t mean anything, even if it’s one value in a spreadsheet column full of other numbers.

The insight, the knowledge, comes when I see “$23.42” and understand its place in a larger dataset as the mean dollar value of my online shoppers in Naperville, IL., a market I’ve done no advertising in but which is outperforming 12 of the 15 markets where I’m spending a ton on ads.

My data didn’t “know” that. My data offered that up to me but it was my putting that number in the context of other factors – including some that aren’t quantifiable – that created the insight and knowledge necessary to make a decision.

I understand that we’ve all been sold the “big data” narrative for the last several years, something that’s only become more pervasive as artificial intelligence is increasingly used in data analysis. Even the best AI, though, won’t be able to provide the same context as an experienced human manager who filters the data through her own knowledge. More troubling is the potential that AI could discount certain things it deems irrelevant but which are actually vitally important, leading to exactly the wrong decision being made.

There are valuable insights to be extracted from raw data, but those insights come from people. The data doesn’t “know” anything, it just presents. Yes, “Know what your data knows” is a catchy tagline because of the symmetry. A more accurate statement would have been something like “Better data, better decisions.”

Chris Thilk is a freelance writer and content strategist who lives in the Chicago suburbs.