For the past three years, the prevailing narrative about AI has been that it would replace workers, eliminate entire job categories, and fundamentally change who businesses need to employ. CEOs repeated it in earnings calls. LinkedIn was full of predictions about which roles would disappear first. The general vibe was that if you didn’t adopt AI immediately, you’d be left behind by the companies that did.
Then the data showed up.
A study published this month by the National Bureau of Economic Research surveyed 6,000 executives across the U.S., UK, Germany, and Australia. The finding: nearly 90% of firms say AI has had no measurable impact on employment or productivity over the past three years. Not a small impact. No impact. Separately, PwC’s 2026 Global CEO Survey of 4,454 executives found that 56% reported neither increased revenue nor decreased costs from AI. Only 12% saw gains in both.
So the tool that was supposed to replace your workforce hasn’t replaced anyone. And for most companies, it hasn’t produced measurable results. But that’s not the interesting part. The interesting part is what’s happening inside the small group of companies where AI is actually working. Because what they’re doing tells you everything about how to think about this technology, whether you run a $500M enterprise or a $5M business trying to grow.
The Replacement Narrative Was Always Wrong
The idea that AI would replace large portions of the workforce was built on a misunderstanding of what the technology actually does well. AI is exceptional at pattern recognition, text generation, data processing, and automating repetitive tasks. It’s terrible at strategy, judgment, nuance, relationship-building, and understanding context that isn’t in the training data. In other words: it’s a very powerful tool that still needs a human who knows what they’re doing to point it in the right direction.
The NBER research actually supports this. An earlier study from the same organization looked at 5,179 customer support agents and found that AI tools increased productivity by about 14% on average. But the distribution was telling. Newer, less experienced workers saw improvements of around 35%. The most experienced, highest-performing workers saw almost no improvement at all. The AI was essentially capturing how the best agents worked and teaching the less experienced agents to mimic those patterns. It was a training tool, not a replacement tool.
IBM’s recent decision to triple their entry-level hiring tells the same story from the opposite direction. They looked at what would happen if AI automated junior-level tasks and realized it would destroy their leadership pipeline. You can’t develop experienced leaders in ten years if you never hire anyone to develop. The humans aren’t optional. They’re the whole point.
Even Daron Acemoglu, the Nobel laureate whose MIT research is often cited in AI productivity discussions, found a modest 0.5% increase in productivity over a ten-year horizon. His take was honest and worth repeating: 0.5% is better than zero, but it’s not the revolution that the tech industry and media have been promising.
What the 12% Got Right
Here’s where the data gets useful. PwC identified a small group of companies (they called them the “vanguard”) that actually reported both cost savings and revenue gains from AI. That group was two to three times more likely to have embedded AI extensively across products, services, demand generation, and strategic decision-making. They had defined roadmaps. They had responsible AI frameworks. They had technology environments built for enterprise-wide integration. Companies with those strong AI foundations were three times more likely to see meaningful financial returns.
Read that list again. Roadmaps. Frameworks. Integration. Strategic decision-making. None of those are technology words. They’re strategy words. The 12% that got results didn’t just adopt AI. They built a strategy around it. They defined what problems they were solving, measured whether the solutions worked, and integrated the technology into how their business actually operates.
The other 88% were mostly running pilots. Testing tools. Experimenting. Which is fine as a starting point, but experimentation without a plan to scale what works is just organized curiosity. It produces knowledge but not results.
Apollo’s chief economist, Torsten Slok, compared this to the IT productivity paradox of the 1980s, when economist Robert Solow noticed that massive investments in computing technology weren’t showing up in productivity data. The problem wasn’t that computers were useless. The problem was that companies were bolting new technology onto old processes without rethinking how they worked. The businesses that eventually rode the IT productivity surge in the 1990s were the ones that reorganized their operations around the technology. Same technology, different approach, dramatically different results.
AI Is a Force Multiplier, Not a Replacement
The most useful way to think about AI, whether you’re running marketing, operations, sales, or any other function, is as a force multiplier. It makes the people who already know what they’re doing significantly more productive. It lets a three-person marketing team produce output that used to require six people. It lets someone who understands strategy move faster from insight to execution. It gives experienced professionals more time for the high-value work by handling the repetitive stuff.
But it doesn’t replace the knowledge. It doesn’t replace the judgment. And it doesn’t replace the strategy.
Give AI to a marketer who understands brand positioning, customer acquisition, and how to connect marketing activity to revenue, and they’ll use it to produce better work faster. They’ll use it to draft content, analyze data, build reports, test messaging, and automate the parts of their workflow that eat time without adding value. That’s a real competitive advantage.
Give the same tool to someone who doesn’t know what good marketing looks like, and you’ll get more bad marketing, faster. You’ll get generic blog posts that sound like every other AI-generated blog post. You’ll get campaigns built on surface-level thinking with professional-sounding copy wrapped around a hollow strategy. The output will look polished. The results won’t change.
This is the gap that the NBER and PwC data is measuring. It’s not an AI gap. It’s a strategy gap. The companies seeing returns have people who know what to do with the tool. The companies that don’t are giving a powerful instrument to people who can’t play it and wondering why the music sounds off.
What This Means for Your Business
If you’re a business owner or marketing leader at a company somewhere between $2M and $20M in revenue, here’s what the AI productivity data actually means for you.
Stop Worrying About Being Replaced. Start Worrying About Being Outpaced.
AI isn’t coming for your job. It’s coming for your inefficiencies. The businesses that figure out how to use it well will operate faster, produce more, and make better decisions than the ones that don’t. The gap won’t be between companies that use AI and companies that don’t. It’ll be between companies that use AI strategically and companies that use it randomly.
The Tool Is Only as Good as the Person Holding It
If your marketing strategy is unclear, AI won’t clarify it. If you can’t articulate who your customer is, what they care about, and why they should choose you, AI will just generate content that’s vaguely aimed at nobody in particular. Very fast. Very professionally. Very useless. The strategy has to come first. Then AI becomes the engine that executes it.
This is why the experienced professionals aren’t threatened by AI. They’re the ones who know what to ask for, how to evaluate the output, and how to integrate it into a plan that actually moves the business. AI doesn’t replace that expertise. It makes it more valuable, because now the person with the strategy can execute at a scale that used to require a much larger team.
Use It Where It Actually Helps
The practical applications of AI for most businesses aren’t the dramatic, headline-grabbing use cases. They’re the quiet productivity gains that compound over time. Using AI to draft first versions of content that a human editor refines. Using it to analyze customer data and surface patterns you wouldn’t have spotted manually. Using it to automate reporting so your team spends less time building spreadsheets and more time acting on what the data says. Using it to speed up research so strategic decisions are grounded in better information.
None of those applications replace a human. All of them make a human more effective. And all of them require a human who knows what “good” looks like to direct the work and evaluate the output.
Don’t Outsource Your Thinking
The biggest risk of AI isn’t that it doesn’t work. It’s that it works just well enough to make people stop thinking. It generates something that looks professional, reads smoothly, and passes the gut check. So the person who requested it approves it without asking whether it’s actually good, whether it serves the strategy, whether it’s saying the right thing to the right audience. The output becomes the standard instead of the standard being applied to the output.
This is how you end up with a marketing operation that’s producing more content than ever and seeing the same results. More activity, same revenue. The volume went up. The thinking didn’t.
The Paradox Resolves the Same Way Every Time
Solow’s paradox from the 1980s eventually resolved. The IT investments that seemed unproductive for years started generating enormous returns in the 1990s and 2000s, once businesses figured out how to actually integrate the technology into their operations. The productivity growth rate increased by about 1.5 percentage points during that period.
AI will likely follow a similar curve. Slok and other economists describe it as a J-curve: flat or declining results early, followed by an accelerating upswing as companies learn how to use the technology effectively. The question isn’t whether AI will produce results. It’s which companies will be positioned to capture those results when the curve turns.
The answer, based on every previous technology wave and on the current data, is the companies that did the strategic work. The ones that built a foundation, defined what they were trying to accomplish, and integrated the tool into a system designed to produce measurable outcomes. Not the ones that adopted the tool first. Not the ones that spent the most. The ones that thought about it most clearly.
If you’re using AI in your business right now (and we’d argue you should be, at least in some capacity), the question to keep asking yourself is simple: is this making us better at executing our strategy, or is this replacing our strategy? If the answer is the former, you’re in the 12%. If the answer is the latter, or if you’re not sure, that’s worth pausing on.
The tool is real. The potential is real. The people who know how to use it will run circles around the people who don’t. And the people who know how to use it are the same people they’ve always been: the ones with a clear strategy, the discipline to execute it, and the judgment to know when a tool is helping and when it’s just generating noise.
AI didn’t change that equation. It just raised the stakes.
Big Brain Strategy helps businesses build marketing systems that use the right tools in service of the right strategy. If you’re producing more marketing than ever and seeing the same results, the problem isn’t the tools. Let’s talk.
bigbrainstrategy.com
The brains behind your growth.
Leave a Reply