The easiest AI business case is written in cost.
How many enquiries can be automated? How many hours can be removed? How much headcount can be avoided?
Those are legitimate questions. But they are not the whole strategy.
A stronger question is what the organisation can now do because routine work has been reduced. Can people solve more valuable problems? Can the business offer a service that was previously too expensive to scale? Can a cost centre become part of a growth engine?
IKEA’s experience with its Billie chatbot is useful precisely because it moves the conversation beyond the technology.
The case has been presented as a chatbot-to-billion-euro growth story and as an example of using AI without losing the human touch. Both capture part of the picture. But the more important lesson is how IKEA combined automation, people and service design to create a different operating model.
In this article we will cover:
- what the IKEA case actually shows
- why efficiency needs a destination
- how reskilling can turn automation into stronger capability
- where this fits with the Invent Group approach
What the IKEA case actually shows
Ingka Group, the largest IKEA retailer, introduced its AI-powered customer service chatbot Billie during its 2021 financial year.
Between 2021 and 2023, Billie resolved approximately 47% of the enquiries it received. That amounted to 3.2 million interactions and nearly €13 million in reported savings.
A narrow efficiency strategy could have stopped there.
Routine work had been automated. Costs had fallen. The resulting capacity could have been translated directly into a headcount target.
Instead, Ingka reskilled 8,500 call-centre co-workers in areas including remote interior design, digital retail sales, relationship-building and the handling of complex customer enquiries.
At the end of the 2022 financial year, sales through Ingka’s remote customer meeting points had reached €1.3 billion, equivalent to 3.3% of total sales. The company said it wanted that proportion to reach 10% over the following years.
The €1.3 billion figure needs to be understood carefully. It represents total sales through the remote channel, not a measured €1.3 billion of incremental revenue directly attributable to Billie.
But that distinction makes the lesson stronger, not weaker.
The chatbot did not create the outcome on its own. The value came from the system built around it: automation for routine demand, trained people for more complex needs, digital channels through which they could serve customers, and a commercial model that turned expertise into a scalable service.
The technology created capacity. The organisation decided what that capacity should become.
Efficiency needs a destination
When AI removes repetitive work, leaders have a choice.
They can treat the capacity created as an immediate cost-saving exercise. Or they can ask where that time, knowledge and experience could create greater value.
Efficiency is not the problem. The problem is efficiency without a destination.
If every successful automation project ends with the same conclusion — that fewer people are required — the organisation may become leaner without becoming more capable. It can lose operational knowledge, customer understanding and the people best placed to recognise the next opportunity.
The wider market reflects this tension. The World Economic Forum found that 77% of surveyed employers planned to reskill or upskill workers in response to AI, while 41% also anticipated workforce reductions as tasks became automated. Almost half expected to move people from roles exposed to AI disruption into other parts of the business.
Those outcomes are not determined by the technology alone. They are strategic choices.
In the IKEA case, the unanswered enquiries were not simply evidence that the chatbot needed to become more powerful. They pointed towards work that required context, judgement, taste, relationship-building and complex problem-solving.
Rather than forcing the technology further into work it was less suited to perform, IKEA moved human capability towards the parts of the customer experience where it could contribute more.
That is a more useful way to think about human-AI collaboration.
AI handles volume, repetition and speed.
People handle ambiguity, trust, interpretation and connection.
The goal is not to preserve every role or process unchanged. Ingka continues to make conventional organisational decisions and announced separately in March 2026 that up to 800 Group Function roles might become redundant as part of a wider simplification programme, alongside plans for 500 new store roles.
The lesson is therefore not that IKEA never reduces headcount.
It is that capacity released by AI does not have to default automatically to headcount reduction. It can be deliberately redirected towards better service, new skills and new sources of value.
From isolated automation to a wider operating model
IKEA’s current AI strategy suggests that this thinking now extends beyond one chatbot.
In a 2026 interview, Ingka Group Chief Digital Officer Parag Parekh described a portfolio organised around three areas: customer experience, supply chain and logistics, and back-office productivity.
AI initiatives are assessed across two further dimensions: whether they are customer-focused or co-worker-focused, and whether they are intended to create growth or reduce cost.
IKEA is not ignoring efficiency. It is using AI to improve fulfilment, logistics and back-office productivity. But cost reduction is one part of the portfolio, rather than the definition of the entire strategy.
Customer growth, co-worker capability and new service experiences sit alongside it.
The customer-facing opportunity is particularly significant.
IKEA says that helping a customer design a space historically required an average of six hours of co-worker time. Through automation, that has been reduced to approximately 30 minutes per room per co-worker, making forms of design support possible at a price point that would previously have been difficult to offer at scale.
That is more than productivity. It changes the economics of the service.
IKEA is also treating adoption as an organisational challenge. By June 2026, it had completed AI literacy programmes for 40,000 co-workers and was running pilots in three countries using store-level ambassadors to identify and demonstrate useful applications.
The principle is simple: a system has not succeeded because it has been built. It succeeds when people understand it, trust it and use it to improve real work.
Where this fits with the Invent Group approach
At Invent Group, we believe every AI efficiency case should have a second line.
The first line describes the effort removed.
The second should explain the capability created.
What will people be able to do better? What customer need can now be served? What decision can be made sooner? What new product, service or commercial opportunity becomes viable?
That means starting with the problem and the desired outcome before choosing the technology.
It also means bringing strategy, technology, commercial thinking, product design and implementation into the same conversation. Reskilling, adoption and service design are not activities to add after the system is deployed. They are part of designing the system properly.
Invent Group’s wider approach is to turn innovation into working capability: stronger products, better systems and more scalable ways of working. Efficiency matters because it improves performance today. Opportunity matters because it creates room for growth tomorrow.
The IKEA case shows what that can look like in practice.
AI reduced routine demand. People developed new skills.
The customer experience expanded.
And the organisation created a route towards value that a headcount-only business case might never have found.
Take home
- AI efficiency creates the greatest value when the capacity released has a deliberate destination.
- The strongest adoption strategies consider customer growth and co-worker capability alongside cost reduction.
- Reskilling, service design and commercial thinking are part of AI implementation, not secondary workforce activities.
- Technology creates the possibility. The operating model determines the outcome.
A practical first step
Choose one workflow where AI could remove a meaningful amount of repetitive work. Before calculating the potential headcount saving, ask:
- what customer needs would still remain unresolved?
- what higher-value work could people do with the capacity created?
- what skills or training would that require?
- could the change enable a better service, product or revenue stream?
- how would success be measured beyond hours or cost removed?
That changes the question from:
“How many roles could this technology replace?”
to:
“What could this organisation now do that it could not do before?”