Multiplier, Not Reducer: What IKEA Understood About AI That Most Companies Get Wrong

Two companies looked at the same technology this year and made opposite decisions.

In March, the fintech company Block cut its workforce from roughly 10,000 people to fewer than 6,000. Its CEO did not hide the reason. “This is not driven by financial difficulty,” (opens in a new tab) Jack Dorsey told staff, “but by the growing capability of AI tools to perform a wider range of tasks.” Around the same time, IKEA’s parent company looked at a chatbot that had quietly taken over nearly half its customer queries (opens in a new tab) and decided not to cut anyone. It retrained 8,500 call-center workers into remote interior-design advisors instead.

Same capability. Opposite instinct. One company used AI to shrink. The other used it to grow. And the difference between those two instincts is going to decide which marketing organizations pull ahead over the next three years.

Here is the argument in one line. AI is not a reason to remove your people. It is a reason to remove the work that buries them. The companies getting the most out of this technology are not the ones with the smallest headcount. They are the ones who used AI to unburden their human intelligence layer so it could finally do the work that only humans can do.

The reducer is betting on the wrong capability

Cutting is the seductive option. It shows up on the next quarterly report. It reads as decisive. And when a competitor announces an AI-driven reduction, the pressure to match it is real.

The mistake is not that these companies believe in AI. They are right to. The technology delivers today, and the gains are real, not someday-theoretical. The mistake is in what they think AI delivers. The reducer assumes the win is substitution: swap a person for an agent and pocket the difference. But substitution is the one move the technology does not make well. A recent Harvard Business Review analysis (opens in a new tab) found that most AI-justified layoffs are being made in anticipation of roles AI will supposedly absorb, not in response to work it has already taken over cleanly. They are not cutting because an agent replaced a person’s judgment. They are cutting because they assume one soon will.

That is a bet on the wrong capability. What AI reliably delivers right now is volume and speed. It strips out manual effort, compresses timelines, and handles production at a scale no team could match by hand. What it does not deliver is strategy, taste, or accountability. Cut your people on the assumption that it does, and you keep the engine while throwing away the hands that know how to aim it.

The honesty in that statement is useful, because it names the bet out loud. Block is wagering that automation can stand in for judgment. The capability is real. What’s misplaced is the belief that it replaces your people instead of amplifying them. And notice what the reducer gives up in that trade.

The multiplier’s bet pays off whether the machine improves or not

IKEA made the other bet. When its “Billie” assistant absorbed close to half of all inbound queries, the company did not treat the remaining work as overhead to be trimmed. It treated it as a signal. The questions customers still wanted a human for were the hard ones. Designing a room. Reading taste. Solving a space that doesn’t quite work.

So IKEA took 8,500 people who already understood the catalog and the customer, and it pointed their expertise at a higher-value problem. Those remote design conversations generated about €1.3 billion in a single fiscal year (opens in a new tab), and the company set a target to grow that share of sales to ten percent by 2028.

There is a fair objection here, and it is worth meeting directly. IKEA is not a pure fairy tale. In 2026 its corporate parent did cut roughly 1,650 office roles (opens in a new tab) under real macroeconomic pressure. But the cuts hit corporate functions, not the reskilled service workforce. That distinction is the whole lesson. The multiplier company is not the one that never trims. It is the one that knows which capability to protect and grow when the technology hands it the choice.

What the evidence actually says about augmentation

Step back from the two stories and the broader data points the same direction. The firms treating AI as a multiplier are not being sentimental. They are being correct.

PwC’s 2026 Global AI Jobs Barometer (opens in a new tab) found that companies most exposed to AI have roughly tripled their lead in productivity growth since 2022, and that roles with high augmentation potential saw job postings rise, not fall. Stanford’s Digital Economy Lab reached a similar conclusion: today’s systems are primarily augmenting human work rather than automating it away (opens in a new tab), with the aggregate employment effect measured in fractions of a percent.

Marketing is a clean example of the pattern. AI adoption across marketing teams reached roughly 91 percent in 2026 (opens in a new tab), up from around 63 percent a year earlier. Yet marketing-manager roles are growing, not shrinking, because the technology automates execution and leaves strategy, judgment, and accountability exactly where they were. The machine got faster at the doing. It got no better at the deciding.

That division is the heart of the multiplier mindset. Agents handle production volume. Humans handle decisions. Get that boundary right and AI stops being a threat to your team and starts being the thing that finally lets them work at the top of their license.

Martech already learned this lesson the hard way

If you work in marketing technology, you have lived a smaller version of this story already.

For a decade the answer to every gap was another tool. The result is the stack most teams now carry: too many platforms, too many logins, and more integration work than anyone can keep up with. The 2026 wave of martech consolidation is, in large part, a reaction to that exhaustion (opens in a new tab). Budgets are tighter, RevOps owns more of the decisions, and teams are done paying for capability they can’t actually operate.

Here is where a lot of organizations are about to repeat their first mistake with a new toy. Bolting AI agents onto a fragmented stack does not reduce the friction. It adds a layer on top of it. Adding AI to an existing workflow is not the same as designing a workflow for AI. The teams seeing real gains are the ones using AI to make their people more effective, not the ones using it to paper over a mess they never cleaned up.

The multiplier is an architecture, not an attitude

This is the part most “AI as a force multiplier” articles skip. They get to the slogan and stop. But unburdening your human layer is not a mindset you adopt in a town hall. It is something you have to build.

The reason most AI initiatives stall has very little to do with the models, which are plenty capable, and almost everything to do with a missing layer between the model and the business. Someone has to frame the right problem. Someone has to harmonize the data the agent acts on. Someone has to set the guardrails and decide when a human takes over. That connective tissue is what we call the AiQ framework (opens in a new tab): three kinds of intelligence working together. Human intelligence frames, designs, and governs. Applied intelligence encodes the workflows and agents that do the work. Platform intelligence runs underneath on the systems you already own.

AiQ Cortex (opens in a new tab) is how we make the applied layer real. It embeds AI orchestration directly into the delivery workflow, and it does so without replacing your platforms or your people. In practice that has meant compressing delivery timelines by five to ten times and taking roughly 90 percent of the manual effort out of routine work. Read that number again with the argument in mind. Ninety percent of the busywork removed is not 90 percent of the team removed. It is 90 percent of the team’s time handed back, redirected from production toward the judgment that actually moves a business.

That is the multiplier made concrete. The reducer automates the human out of the loop. The multiplier keeps the human in command and clears the friction underneath them. Same technology. The architecture is what decides which one you get.

So the question is no longer whether AI will change your team. It will. The question is whether you will use it to make your team smaller, or to make the people you already trust dramatically more powerful. One of those bets compounds. The other one hits a floor the moment the cutting stops.

Related Page

Find out how AI can transform your team

If you are trying to figure out where AI can lift the load off your team instead of replacing it, that is exactly the conversation our AiQ team has every day.

Get Started