About this article:
Still Thinking is a personal reflection on what marketing leaders have chosen to keep doing themselves in an AI-everything world. The habits, rituals, and creative practices they are deliberately protecting.
Last month marked my 23rd year with Google. So the following statement may seem unusual considering I’ve spent most of my adult life working for one of the largest tech companies on the planet.
The secret to successfully unlocking the power of AI is human decision-making.
I spend a great deal of time during my day analyzing data, examining analytical output, reviewing reports and making critical decisions in order to help some of the world’s largest brands and agencies accelerate growth. Additionally, I actively participate in industry conversation, sharing my thoughts around AI, the transformation of retail and the future of marketing.
The benefits we receive from leveraging AI and LLMs in our everyday lives are undeniable. We can summarize complex amounts of data, search for patterns within that data and answer with near certainty, a wide variety of questions that would normally take people hundreds of hours to accomplish manually.
However, as AI continues to advance exponentially, I find myself questioning less of what I expect technology to provide and more of what human aspects of my job are most important to preserve.
The function that I'm most determined to protect and continue to perform myself is critical thinking.
And it isn’t because I feel that AI doesn’t have intelligence. AI is wicked smart. It's because the easier it gets to obtain answers, the greater the risk we stop asking questions.
This is nothing new. For centuries, every generation has faced a version of this issue. When new technologies emerge, friction is removed. As barriers and effort minimize, the quality of life increases. However, as our quality of life improves, there are trade-offs.
When we stop practicing certain skills, they tend to atrophy. Take GPS for example. This has made navigation so simple (and added magical features like predicting traffic patterns to cut down drive time), but I've lived in northern New Jersey for 10+ years and I couldn't give you directions from my house to Route 46 if my life depended on it. The same can be said for search engines. I sometimes struggle to articulate the why, but I take serious issue with my kids’ assertion that they don't need to memorize anything because they can just Google it.
AI presents a similar challenge – but the skill at risk isn't memorization or navigation: it’s judgement.
It’s easy to assume data you’ve been given is correct when an analysis is well presented and you have pages of supported information. But experience has taught me that some of the most important decisions happen after the analysis is complete.
A few years ago, my team was working on a project for a mattress retailer. Our objective was to assess how much foot traffic a media campaign could drive in a specific geographic region. The methodology was sophisticated. The data was extensive. The outputs looked polished and persuasive. But there was one flaw. The projected store visits exceeded the population of the region.
At first glance, nothing appeared wrong. The analysis checked all the boxes. The numbers were presented confidently. The recommendation looked credible. But, it unfortunately failed a simple test that no AI could decipher: A good ole fashion gut check. Did it make sense?
The answer, obviously, was an emphatic “no”.
That moment has stuck with me as I feel it clearly illustrates a significant challenge we’ll face more and more in an AI driven world. As this technology advances further, output will look progressively more sophisticated, extremely persuasive and harder to dispute. Mistakes won't always be obvious. In fact, many of them will look entirely reasonable.
That's why I believe the skills we often associate with a liberal arts education are becoming more valuable, not less. The ability to think critically. To evaluate competing ideas. To understand context. To recognize faulty assumptions. To ask better questions.
Historically, these abilities were often seen as secondary to technical competencies. Today, I view them as complimentary (if not paramount). Technical tools allow us to quickly attain answers. Critical thinking enables us to determine if we should actually trust those answers.
I also believe that the principles above apply equally as well to Sales which is a field that’s undervalued.
The most effective sales professionals do not achieve success by learning scripts. They achieve success because they know how to listen. They understand human behavior. They can recognize nuances and adjust their approach based on these micromoments. They know when something feels wrong even though the data indicates otherwise.
In many ways, that's the same muscle we're going to need as AI continues to integrate itself into our daily activities and our workplaces. Technology can generate recommendations but it can’t assume responsibility for judgment. That responsibility remains with us.
So in an AI-everything world, the habit I'm deliberately protecting isn't a productivity ritual or a creative process. It’s giving space to slow down long enough to ask pointed questions. Or better yet, “trust, but verify.” Because the greatest risk isn't that AI will start thinking for us, it will be that we’ve become so comfortable with its responses, that we stop thinking for ourselves. And should that happen, we'll lose the very skill that makes all of this technology valuable in the first place.