In the middle of a meeting with a client, an experienced key account manager can’t recall a figure he used to know by heart. What happens to the knowledge we’ve accumulated over an entire career when we gradually hand it over to a tool, and when do we realize it’s missing?
Last week, I accompanied a key account manager to a client meeting. We were in the middle of discussing adjustments to a policy when he couldn’t remember the deductible schedule. Three years ago, he could have recited it in his sleep, he’d told me that himself once. He reached for his tablet, typed the question into his AI tool, and waited while the client looked at him.
No one held it against him. The customer waited politely, as one does. But in those ten seconds, I sensed something that has been on my mind ever since: a slight sense of vulnerability, the feeling of having given something of myself away without realizing it.
Behind this lies a question that goes far beyond this one appointment: If the knowledge that once defined a profession is outsourced to a tool, who does it actually belong to anymore?
Outsourced thinking
This is what the term “cognitive offloading”: the transfer of mental work to a tool that performs a task more reliably or quickly than one’s own mind. Anyone who reflexively passes a question on to an AI tool instead of thinking it through themselves no longer trains their own ability to do so. And what isn’t trained withers away, slowly, imperceptibly, until it’s missing at the crucial moment.
The phenomenon itself is not new. Studies have shown for years that people who rely on navigation systems lose their ability to orient themselves in space. The same thing happened with mental arithmetic when calculators were introduced, and with spelling when auto-correct was introduced. What’s new is the scope. In addition to calculations and directions, an AI assistant can answer almost any question that comes up in a broker’s day-to-day work. If this reliance extends to the entire profession, the loss of these skills will grow as well, especially if there’s no training to counteract it.
The figures show the pace of this development, not the caution involved. In a recent study by the Heute und Morgen Institute of 284 real estate agents in Germany, four out of ten already use AI on a regular basis. More than seven out of ten see it primarily as an opportunity. What the study doesn’t measure concerns me more: how often a response is still reviewed before it’s sent out. And how often it’s simply accepted because the tool has already provided it.
What’s missing from the advisory process
For the insurance industry, this is not just a theoretical problem. An advisor who immediately passes on every inquiry will, over time, lose the intuition for the bigger picture, the very thing that makes for good advice. What this means is knowing why a particular policy suits one client but not another, anticipating which question will come next, and recognizing when a gut feeling is more reliable than any database query. This intuition doesn’t come from looking up answers, but from asking questions you’ve thought through yourself, even the wrong ones.
The same applies to risk assessment in the application process. A system can provide a recommendation in seconds, but the ability to question a recommendation when something doesn’t add up remains a human one. It only develops when you’ve tried to make an assessment on your own without any tools, and have been wrong at least once in the process.
I see this most clearly when training new consultants. Anyone who knows only the ready-made answer from the start has never learned how to arrive at it. In a conversation with a client, this makes a difference you can sense, even if you can’t put your finger on it right away: whether someone understands an answer or is just reading it aloud.
A quick exercise
I don’t think the solution is to put the tools away. I use them every day, and they help me improve in many ways. But I’ve started to consciously give myself a break when it comes to simple questions before reaching for my tablet. It’s not a major decision, more of a small exercise, because I still want to trust my own mind to do what it’s capable of.
In the teams I coach, this pattern often shows up first among the most experienced members. Those who know the most have the most to lose when they stop drawing on that knowledge themselves. Younger colleagues have grown up with this tool; for them, a different question arises: what they’ve never built up on their own because they’ve never had to.
Both deserve attention, but the answer remains the same: consciously keep your own thinking active, just as you would exercise a muscle you want to continue using. In the training sessions I lead, I’ve therefore started incorporating exercises that require no tools—short case studies that participants first work through in their heads before turning to an assistant. The difference isn’t immediately apparent. It becomes evident weeks later, in real-life conversations, when the tool fails or there’s simply no time to use it.
I’m a firm believer in AI tools. I consider them one of the greatest opportunities our industry has had in years. But I don’t want them to take away my own judgment without me even realizing it. I’d rather decide that for myself, every single day.
By Hand
After the meeting, I asked the key account manager to write down the deductible schedule again by hand, without a tablet, from memory, as best he could. He got two numbers wrong. The next day, he tried again. This time, both were correct.
Marcus Selzer
See also: The Hand That Wouldn’t Let Itself Be Washed
Search:
Sponsors:
