"Do more with less" is the phrase almost every executive reaches for when they talk about AI. It sounds responsible. It sounds disciplined. It signals to the board that you're a careful steward of capital.
It's also a ceiling.
When you frame AI as a way to do more with less, you've already decided the shape of the prize before you've looked at it. Less headcount. Less cost. Less spend per unit of output. The numerator (what you actually produce, sell, and win) stays roughly fixed, and the whole exercise becomes a hunt for a smaller denominator. It's a small question to ask of a very big tool.
The companies that will pull away over the next few years are running a different equation. They're using AI to do more with more: more capability, more output, more reach, aimed at taking ground their competitors can't. That mindset is harder. It's more complex, more uncertain, and far more demanding of the person at the top. Which is exactly why most companies avoid it, and why the ones who don't will win.
Small thinking is the default, and the data shows it
This isn't a strawman. The efficiency framing dominates real AI strategy today.
In the EY-Parthenon Growth Survey, 63% of companies said they use AI primarily for efficiency and productivity. Only 14% use it to stay ahead of competitors, and just 7% to diversify revenue streams. Meanwhile, 78% of executives say they believe AI will accelerate their organization's growth rate. Read those numbers together... most leaders believe AI is a growth engine, and most are using it as a cost lever anyway.
McKinsey's 2025 State of AI report tells the same story from another angle. Eighty percent of organizations set efficiency as an objective of their AI initiatives. Yet only 39% can point to any measurable EBIT impact at the enterprise level, and among those that can, most put the effect below 5%.
That's the tell. When you point the most consequential technology shift in a generation at "shave a few points off our cost base," a few points off your cost base is roughly what you get. The tool delivers exactly the size of the question you asked it.
Doing more with less and doing more with more are not the same bet
Here's the part the efficiency crowd misses: the most valuable AI work isn't a milder version of cost-cutting. It's a different bet with a different payoff curve.
McKinsey found that the companies seeing the greatest value don't drop efficiency. They add growth and innovation objectives on top of it. High performers are significantly more likely than the rest to set growth or innovation as explicit AI goals, and the difference shows up where it counts: competitive differentiation, customer satisfaction, and gains in market share that efficiency-only strategies simply don't produce.
The size of the prize isn't a rounding difference. BCG's research on AI leaders versus laggards found leaders achieving 1.7x the revenue growth, 3.6x the three-year total shareholder return, and 1.6x the EBIT margin. And here's the crucial point: those same leaders also captured about 40% greater cost reductions.
The leaders weren't choosing between offense and defense. They were better at both... The "do more with less" framing quietly assumes a tradeoff: that you spend your AI budget either on efficiency or on growth. The actual leaders refused that tradeoff. They treated efficiency as table stakes and growth as the point. BCG's broader research on transformation is direct about why: the strongest programs carry a dual mandate, cutting cost and growing revenue at once, because cost-cutting in isolation tends to leave an organization without the direction or momentum to actually compete.
Look at where the money actually goes
If you want the clearest evidence that "more with more" is the smarter play, look at what the most thoughtful companies do with the productivity AI gives them.
The fear is that AI productivity translates directly into layoffs. The data says otherwise. In EY's December 2025 AI Pulse Survey, only 17% of companies said AI-driven productivity gains led to reduced headcount. The rest were plowing those gains back into building the business: 47% into expanding existing AI capabilities, 42% into developing new ones, 41% into cybersecurity, 39% into R&D, and 38% into upskilling and reskilling their people.
EY's summary of that behavior is the sharpest version of the whole argument: these companies "aren't trying to run the same race with fewer people; they are effectively buying the capacity to run a faster, more complex race."
That's the difference in one sentence. Small thinking takes the capacity AI frees up and hands it back to shareholders as a one-time saving. Big thinking takes that same capacity and spends it on a bigger race: more products, more markets, more of the things only humans-plus-AI can now do. One is a cash-out. The other is a compounding investment in winning.
Why most companies still think small
If doing more with more is so clearly the better bet, why do so few play it?
Because it's harder, and the difficulty is mostly about leadership, not technology.
Cost-cutting is legible. You can put a number on it, socialize it to the board, and book it this quarter. It requires no vision about where the market is going, no conviction about a future that doesn't exist yet, and no willingness to absorb the disruption of becoming a different company. You can do it badly and still look responsible. "Do more with less" is the answer you give when you don't have a more ambitious one.
Growth is the opposite. It's ambiguous. The payoff is further out and harder to attribute. It demands that you redesign workflows, reorganize teams, and ask people to do genuinely new work, not just the old work faster. That's why so much AI investment stalls. The MIT NANDA report, The GenAI Divide, found that roughly 95% of enterprise generative-AI pilots failed to deliver a measurable financial return, and its authors were explicit that the failures were about how organizations deployed the technology, not the quality of the models. The winners concentrated on a few high-impact uses; the losers spread AI thinly across everything and hoped.
Thin, timid, scattered deployment is what small thinking looks like in practice. It's not a separate failure mode from "do more with less." It is that mindset, met by reality.
This is a leadership test, not a technology test
Here's the uncomfortable conclusion the evidence points to.
If you're a weak leader, "do more with less" will look like wisdom. It's controllable. It protects you. It lets you claim progress without betting on a future you'd have to defend. You can hide behind it.
If you're a strong leader, the goal isn't a leaner version of the company you already have. It's a bigger one. One that uses AI to do things your competitors can't yet do, to enter markets you couldn't previously serve, to take ground while everyone else is busy trimming. That ambition is heavier to carry. Bain's research found that more than 80% of CEOs are dissatisfied with their AI transformation's progress, and that most falter not on strategy but on execution: translating ambition into scaled outcomes with speed and consistency. The companies that succeed tend to have a leader with genuine conviction about how AI changes the business, not a delegated cost program.
Doing more with more is more complex by definition. More moving parts, more change for people to absorb, more places to be wrong. It requires holding a clear picture of the outcome you're driving toward and keeping the organization pointed at it through the mess. That's the work. It doesn't delegate.
Start from the outcome, not the budget line
This is where thinking backward earns its name. Small thinking starts from the budget line: here's our cost base, where can AI shave it down? Every answer it produces is bounded by the question.
Big thinking starts from the outcome: what would it take to win our market outright over the next three years, then works backward to the capabilities, the workflows, and only at the very end, the tools. When you start there, AI stops being a cost lever and becomes what it actually is: the largest expansion of organizational capacity most of us will see in our careers.
You can spend that capacity on a smaller race. Or you can spend it conquering your industry... The technology is identical. The mindset is not, and neither are the results.
Stop thinking small.