#45 - Use it or lose it: your brain in the age of AI
Part 1 of 2: How to build and protect your strategic thinking
At CPO Connect last week, 70+ senior product leaders gathered at AKQA to do what product leaders do best: identify problems and solve them. I co-facilitated two breakout sessions with my friend James, exploring which cognitive skills are most at risk from over-reliance on AI. Across both groups — somewhere between 25 and 30 people in total — the same skills kept coming up: strategic thinking and decision making. The concern, though, was almost entirely directed at someone else.
Early-career professionals are outsourcing the effortful thinking before they've had the chance to build it. Writing a deck isn't just about creating pretty slides — it's how you learn to structure an argument. Sitting with a hard problem is how you build judgement. If you outsource the whole task, you lose the skill underneath. The experienced leaders in the room understood this intuitively, but what they were less sure about was whether it applied to them.
The research suggests it does. Use it or lose it applies to the brain as much as it does to muscle — and AI is making it very easy not to use it. Kosmyna et al (2025) found that heavy AI users showed the lowest brain engagement of all groups and got measurably lazier with each session. The researchers called it cognitive debt, and its impact on memory was immediate — participants struggled to recall content they'd produced just minutes earlier (something noted in our discussion too). A separate study by Microsoft Research and Carnegie Mellon University (2025) surveyed 319 knowledge workers and found that for 40% of tasks, participants reported using no critical thinking whatsoever. The more confidence workers placed in AI to complete a task, the less they engaged their own judgement. The researchers described it as losing the muscle memory built from doing the work yourself.
As a former Product Director now retraining in organisational psychology, I've experienced this first-hand. When I over-rely on Claude, I notice it — the thinking feels shallower, the ideas less mine. So I went back to a framework I built a few years ago: a core competencies model structured around three relationships — to yourself, to others, and to your context. Strategic thinking and decision making both sit in that third category. I dusted it off on Friday and updated it for this piece.
This is the first of a two-part series. Here, we go deep on strategic thinking, with part two covering decision making. I see them as two sides of the same coin — you need both to act well under pressure and uncertainty. For each competency we’ll look at the cognitive skills underneath it, how they develop across the lifespan, explore exercises to build them, and share a practical guide for when to use AI — and when to do the thinking yourself.
What is strategic thinking?
Let’s start with what is strategy. I really like Mark Cuban’s definition on Masterclass: it’s how you’re going to win. All the frameworks and matrices that we get bogged down in are simply mental scaffolds — not the strategy itself — to get you closer to that answer.
So strategic thinking refers to the abilities that enable you to create a strategy. And you’re creating strategies for life events all the time: moving home, thinking about your financial future (e.g. FIRE), deciding which school to send your kids to. You just don’t always see it as strategic thinking because it didn’t involve a lengthy presentation to the Board (just a discussion with your family!).
Here's what those underlying skills actually are:
Visioning: Define a compelling and coherent picture of a desired future state that guides direction and decision making.
Goal setting: Translate that vision into specific, measurable outcomes that focus effort and track progress.
Planning: Sequence actions, resources, and timelines to move from current state to desired outcome.
Systems thinking: Identify how multiple variables interact and influence each other across different levels, recognising that changing one element produces effects throughout the whole.
First principles thinking: Strip a problem back to its foundational truths and reason upward from there, rather than by analogy or inherited assumption.
Root cause analysis: Trace a problem back to its originating cause rather than its visible symptoms, distinguishing between what is presenting and what is actually driving it.
How these skills develop
Strategic thinking is a distinctly human capability — it depends on the prefrontal cortex, the most recently evolved part of the brain, and the seat of what psychologists call executive function. It's what separates us from other animals: the ability to think beyond the immediate moment, simulate future outcomes, and act on reasoning rather than instinct.
These capacities don't arrive fully formed, and they don't stop developing on any fixed schedule. The oft-cited claim that the brain stops maturing at 25 turns out to be an artefact of the original research — the oldest participants were in their early twenties, so 25 was an extrapolation, not a finding. A 2025 University of Cambridge study mapped the brains of 3,802 people aged zero to ninety and identified five distinct phases of brain development across a lifetime. The adolescent phase of neural wiring extends to around 32, on average — at which point the brain shifts into its longest era, lasting over three decades. Development doesn't stop there either. It continues to be shaped by experience, challenge, and significant life events throughout adulthood (what Zittoun calls ruptures).
So how do we develop our strategic thinking?
Through sitting with hard problems, like deciding whether to take a job or not. In this example, you start with imagining different scenarios — what it would be like to accept and do that role, versus if you didn’t take it. How you might make childcare work with a longer commute to be in the office 3 days a week. Who you could call on for additional support. Identifying the pros and cons, the benefits and the costs (weighing them up engages critical thinking, which is under the decision making competency). Evaluating if taking a higher salary for a couple of years is worth it to get you closer to early retirement, and what is the risk of burning out.
Doing this requires embracing discomfort. The cognitive effort of mapping out possibilities creates mental friction as your brain engages across different neural pathways: memory and imagination, emotion and reasoning, attention and inhibition. This friction is an essential part of the process, to internalise the knowledge and take ownership of the decision this ultimately leads to — something the CPO Connect group noted. So now let's get practical.
Exercises to develop your strategic thinking
Here are three exercises you can use yourself or share with your team to develop strategic thinking.
Exercise 1: What if (and then what?)
Pick a trend, a constraint, or an assumption in your work or life and ask: what if this changed? e.g. what if remote work became the norm again? What if your biggest competitor halved their prices? What if you got the promotion?
Then follow it through:
First order effects: what happens immediately as a direct result?
Second order effects: what does that trigger in the people, systems, or markets around it?
Third order effects: what does that change about the broader landscape over time? (use different timelines such as 1 year, 5 years, 10 years, 50 years).
Most people stop at first order, but the strategic thinking happens in the second and third. The further out you can reason with confidence, the stronger your visioning and systems thinking become.
Exercise 2: Work backwards from winning
Sir John Whitmore’s insight in Coaching for Performance was that most people jump straight to performance goals — the metrics and milestones (aka OKRs) — without ever establishing the bigger goal underneath. Goals should be inspiring and aligned with personal values: the measurability matters less than whether the goal actually means something to you.
So before you set a goal, define what winning looks like. Not the output, but the outcome: what changes, for whom, and why that matters to you specifically. For example, my 15 year goal is to support a team of space psychologists to land humans successfully on Mars. This matters because getting there will further our understanding of the universe, and change how we see ourselves (as a multi-planetary species). I see myself as an explorer and this is the ultimate expression of that.
From this you can use a technique called backcasting: take that goal and work backwards, breaking it down into a series of steps or milestones. These can then be measured to ensure you’re making progress and on track. To build on my example: BSc psychology complete > MSc organisational psychology loading this autumn > original research in space psychology; developing contacts in the space sector > mentoring on next cohort of UK Space Agency’s accelerator programme > learning how the space sector works.
Exercise 3: Five whys
A favourite with young children everywhere. Take a problem you’re currently facing — professional or personal — and ask why five times. This enables you to dig beyond the surface level, and get to the foundational elements driving that behaviour.
For example: I’m not making progress on writing my book.
Why? Because I’m not writing consistently.
Why? Because I keep deprioritising it when other work comes in.
Why? Because I don’t protect my time for it.
Why? Because I haven’t committed to it as a non-negotiable.
Why? Because I don’t fully believe I’m capable of writing a whole book.
That last answer is where the real work is. Root cause analysis isn’t about finding the practical fix — it’s about getting to the belief or structural problem underneath. That’s where strategic thinking and change actually begin.
Now let's look at how to integrate AI into your strategic thinking practice — without letting it do all the thinking for you.
AI-augmented strategic thinking
There's a principle hiding in plain sight in the name: generative AI helps you generate more possibilities. It's not there to do the thinking from first principles for you — it's there to extend what you already came up with and surface the blind spots you couldn't see from where you were standing.
The most useful framework I've come across for this is the human-AI sandwich, developed by Beth Kanter (which I first came across in my friend Athena's forthcoming book AI-Q: The new leadership imperative for the human + AI era). The principle is simple: human thinking goes first (bottom bread slice), AI augments in the middle (the filling), and human judgement closes it out (top bread slice). Before you go to AI, assemble what you already know — your existing understanding or research, your assumptions, your rough thinking. Then use AI to expand, challenge or accelerate that. Then come back and apply your own critical eye to what it produced. What's missing? What assumptions are baked in? What would make you reject this entirely? The sandwich only works in that order: starting with AI skips the step that builds the skill.
Applying to the skills listed earlier, here’s what I recommend:
Visioning and goal setting: Define your own vision and goals before you go anywhere near AI. What are you trying to achieve and why does it matter? Then ask the LLM to challenge the coherence — what's missing, what's contradictory, what assumption is doing too much work?
First principles and systems thinking: Identify the key variables and first, second and third order effects yourself first. Then prompt the LLM to surface scenarios and dependencies you haven't considered.
Planning and root cause analysis: Draft your own sequence of steps or your initial diagnosis of what's driving a problem. Then ask the LLM to stress-test it — what have you assumed away, what are the failure points?
Final synthesis: Once you've worked through the above, ask the LLM to summarise the exercise. That summary becomes the basis of your one-pager — but you wrote the thinking it's drawing on.
The temptation with the accessibility of AI is to jump straight into it to do the heavy lifting, so this is going to require discipline not to succumb. The litmus test is going to be can you stand up in front of an exec team and defend your work, because you know it inside-out? If you can’t honestly answer yes, then you have outsourced too much thinking to AI. And if someone hands you a fifty page document, the first challenge is asking them to summarise it for you — verbally. Very quickly you will learn if they’ve internalised that knowledge or are merely a conduit.
One other point worth making: the CPO Connect discussion rightly highlighted that AI is genuinely transformative for many neurodivergent people — helping with tasks where the barrier isn't the thinking, but the execution. Rather than seeing this as cognitive offloading leading to potential skills atrophy, it’s about creating access where structural barriers exist that prevent participation.
If we’re using LLMs to augment human thinking instead of replacing it, that implies it's not actually going to save time and lead to the productivity gains espoused by the AI-bros (where the time saving comes from not doing the work yourself). And several studies have shown this is what is playing out: A 2024 MIT study by Nobel laureate Daron Acemoglu found only a 0.5% increase in [overall economic] productivity over the next decade from AI. Acemoglu himself described the finding as “disappointing relative to the promises that people in the industry and in tech journalism are making.”
Instead, the real value in LLM use is about generating more possibilities and addressing blind spots — which makes the thinking more robust. And in a world where there is increasingly more noise than signal, understanding what to pay attention to and act on will become even more critical.
If you found this useful, here's how I can help you more:
Develop your career as a leader with me through leadership coaching.
Inspire your audience to change their beliefs through In-company keynotes and workshops.
Join the waitlist for my new book, Beliefocracy: How our beliefs govern us.
Commission a piece of novel research.