Buying AI and Using It Well Are Two Different Things

Adoption headlines suggest a technology sweeping through the economy. The harder question is whether companies are building anything new with it, because that is what actually shows up in growth.


Here is a number you have probably seen some version of: roughly one in five firms and businesses are now adopting AI, according to Census data. It sounds like a technology spreading fast. And in a sense, it is.


But adoption figures answer the easy question, the one about whether companies have brought the tool in the door. They say nothing about the harder and more important question, which is what those companies are actually doing with it once it arrives. That second question is the one I care about, because it is the one that decides whether all this activity adds up to a stronger economy.

A tale of two adopters

Imagine two companies that both report adopting AI. They look identical in the statistics.


The first uses AI to trim expenses. It automates a few routine tasks, speeds up some paperwork, shaves cost off processes it already runs. Useful. Real money saved. The second uses AI to build something that did not exist before, a new product, a new service, a genuinely redesigned way of working.


Both count as adopters. Only one is innovating. And the gap between them is where the entire economic story lives.

What the surveys actually show

A recent Deloitte survey dug into exactly this, looking past whether firms had invested in AI to examine what they were investing in it for. The result is worth pausing on. Only about 30 percent of adopting firms are using those investments to create new products and new processes. Everyone else is in the other camp, squeezing more efficiency out of work they were already doing.


So when you read that one in five firms have adopted AI, the more revealing figure is hidden underneath: of those adopters, fewer than a third are using it in the way that drives real innovation. The headline describes a wave. The detail describes a much narrower channel within it.

Why the difference reaches the whole economy

You might reasonably ask why this matters beyond the individual company. If a firm saves money cutting costs, good for that firm. The catch is that trimming expenses by itself does not raise the productivity of the wider economy in the way that inventing new products and processes does.


The concept economists use to track this is total factor productivity. Strip away the contribution of simply adding more workers or more machines, and what is left, the part that comes from combining those inputs more cleverly, is total factor productivity. It is the closest thing we have to a measure of genuine ingenuity in an economy. It rises when we learn to produce more or better things from what we have, which is precisely what new products and processes accomplish. Doing the same old things a little cheaper does not move it.


That is why the adoption versus innovation distinction is not academic hair splitting. A firm cutting costs with AI has bought a better tool. A firm inventing with AI has changed what it makes and how. The first shows up on one company's balance sheet. The second is the kind of change that, in time, registers as a genuinely more productive economy.

Why so many firms stop at cost cutting

If innovation is the prize, why are most firms not chasing it? Because it is genuinely hard. Cutting costs with a new tool is comparatively easy, since it slots into work a company already understands. Inventing new products or rebuilding core processes asks for experimentation, organizational change, new skills, and a tolerance for things not working the first time.


This is also why I would not read the current 30 percent as a final score. Every general purpose technology in history started the same way, with the easy uses first and the transformative ones arriving only once people figured out what the tool could really do. The share of firms doing the harder, more creative work tends to climb over time. Whether it climbs this time, and how fast, is the open question.

The number I am watching

All of which is why, as the data on AI keep coming in, I am watching one figure above the rest. Not the adoption rate. The share of adopters who move past experimenting and cost cutting toward actually creating new products and processes, the roughly 30 percent in the Deloitte data today.


If that share rises, it would tell us this wave of AI investment is quietly laying the foundation for durable growth and the kind of productivity gains that genuinely strengthen an economy. If it stalls, we may end up with a lot of leaner companies and not much more. The difference between those two futures is the thing I will be tracking, and it is a far better gauge of this moment than any adoption headline.


Adriana Kugler, Ph.D., is a labor economist and professor at Georgetown University.

Comments

Post a Comment

Popular posts from this blog

The Grocery Receipt Tells a Different Story Than the Inflation Report

Prices are cooling in the data, and the public is more worried than ever. The contradiction dissolves the moment you stop treating inflation as one thing. Right now, in 2025 and into 2026, the price pressure people feel most is back where the whole episode started: food and energy, with goods alongside. The cause this time is geopolitical rather than pandemic logistics . Tariffs have added up to 0.7 percentage points to goods prices over the past year and a half. The war involving Iran has pushed energy costs up, since roughly 20 percent of the world's oil moves through the Strait of Hormuz, which has been under blockade for several months. And because somewhere between a quarter and a third of globally traded fertilizer passes through that same strait, food prices have climbed too. These are the prices households meet weekly and sometimes daily, at the pump and the checkout, which is why this latest pressure lands so hard on the public mood. And that brings up a puzzle worth sitti...
Read more

A Supply Shock Has Two Acts. We Keep Watching Only the First One

Higher prices are the headline. The slow erosion of demand that follows is the part that actually decides how rough the landing gets. Picture a negative supply shock the way most coverage frames it. Costs jump, growth gets pinched, and the economy ends up in that uncomfortable spot where prices climb while activity drags. That framing is not wrong. It is just incomplete, and the missing half happens to be the half that matters more for where growth goes next. Having spent years studying economies at very different stages of development and having sat on the Federal Open Market Committee as these questions played out in real time, the pattern I keep coming back to is this: A supply shock is not one shock. It is a first shock that can summon a second one. The initial blow lands on supply and shows up as inflation. Then, with a lag, it can reappear on the other side of the economy as a negative demand shock. And that delayed second act is usually where a real slowdown takes root. The ear...
Read more

Why Cheaper AI Could Mean More of It, Not Less

Making a powerful technology cheaper does not always shrink demand for it. Often it does the opposite, and that pattern is worth keeping in mind as the cost of running AI falls. Most coverage of artificial intelligence focuses on what the technology can do. The figure I keep coming back to is a quieter one: what it costs to run. That number, more than any single capability, will shape how far AI spreads and how much it ultimately changes the economy. The number behind the question There are estimates suggesting that the cost of processing for a model with one trillion parameters could fall by roughly 90 percent by 2030. That is not a modest discount. It is the price of a powerful capability dropping to a tenth of what it is today. The instinct, when a price falls like that, is to assume spending on it falls too. Cheaper AI, lower AI bills, end of story. But that instinct is usually wrong, and the reason has a name. Efficiency does not lead to thrift It is called the Jevons paradox: im...
Read more