It’s Not the Tools That Matter: Sense-Making in the Age of AI
Why strategy needs fewer shortcuts, not faster ones.
The AI Gold Rush
If you work in digital strategy, design, or marketing, it’s hard to avoid the rush. AI is everywhere — powering creative, writing content, drafting personas, shaping decks. The general theme I’m hearing everywhere is generally the same: do it faster, cheaper, better.
There’s no doubt AI can be incredibly useful in these contexts, impressive even. Large Language Models already do a solid job of mimicking tone, structuring insight and produce outputs that look and sound the part. But as this broad industry speeds toward automation, a quieter question nags at me: are we actually doing better work — or just doing more of it, faster?
Deliverables Over Dialogue
Much of what passes for strategy today is output-driven. Decks. Diagrams. Canvases. We’ve built an industry on deliverables that signal progress, but often avoid the harder work of sense-making.
As an industry we love a framework — Strategyzer, value prop canvases, customer journey maps and grids. Don’t get me wrong, these can all be useful in the right hands and context. But often in practice they also come with a risk: the illusion of understanding. We fill out the boxes, name the personas, map the touchpoints - and it all looks like strategy, but it doesn’t really feel like a credible one.
Frameworks can help structure thought. But when we lean too heavily on them, they can actually become a way of avoiding thought - compressing complex, emotional, often contradictory realities into something tidy enough for a slide.
AI Is That Trap on Steroids
Now take everything frameworks do and add speed, scale, and surface polish.
LLMs don’t just help you fill in the canvas. They write the value proposition, define the target audience, summarise the research, and spit out the insights — all in perfect slide-friendly prose. They’ll get you to a good enough answer in minutes.
But here’s the problem: in the wrong hands, AI gets you to mediocrity faster than ever. From a strategy perspective, my worry is it doesn’t really help you confront complexity. It doesn’t challenge assumptions (unless pushed). It doesn’t hold tension, or hold you to account. It gives you a finished-looking thing - the appearance of thinking - without the actual discomfort of having done it. If frameworks offered a shortcut through ambiguity, AI is that shortcut turbocharged. What’s the end result? More polished decks, campaigns, websites, apps solving the wrong problems but faster.
Resist the Urge to Outsource Problem Definition
We need to stop outsourcing the hard bit.
The most valuable work in strategy, for me, isn’t the output - it’s sitting with the messy input. It’s spotting the tension no one wants to name. It’s resisting the temptation to resolve ambiguity too quickly. It’s understanding the human structures that shape a business and its priorities - how power, emotion, politics, and misaligned incentives shape the real problem.
AI doesn’t do that. It can’t do that. It doesn’t sense when something’s off, or know when to push back on a brief. It doesn’t know how to read a room. And it certainly doesn’t understand when everyone’s nodding at a plan they don’t actually believe in.
That’s human work. That’s our work.
So What Now?
This doesn’t mean rejecting AI. The opposite, actually. We should use it — but with care. With clarity about what it’s good at, and where we still need to show up and do the thinking.
Here’s what that might look like:
Use AI to summarise, not synthesise.
Use it to stretch your perspective, not set your direction.
Use it to prototype a thought, not finalise it.
Use frameworks to open up discussion, not wrap it up.
And most of all: don’t confuse outputs for outcomes.
What we need now isn’t faster strategy. It’s more critical strategy, strategy that asks harder questions and that pushes back on comfortable answers. That makes space for complexity and builds clarity not from templates, but from real thinking.
In the near future the strategist who can hold the problem longer, ask better questions, and align others around what really matters won’t be automated - they’ll be essential.
