AI Isn't Only for Saving on People. It Frees Them to Add More Value
Image: illustrative — IKEA Billie chatbot and remote design advice
AI Isn’t Only for Saving on People. It Frees Them to Add More Value
What IKEA did with 8,500 service agents is a lesson most automation projects still haven’t learned.
Almost every AI project in customer service is presented the same way: deflected contacts, cost per contact, positions no longer needed.
That’s the easy part. And it’s the part that’s worth the least.
IKEA shows it with three figures that are worth reading together:
- 📊 €13 million saved by its AI assistant in its first two years.
- 📊 8,500 agents who were not laid off, but reskilled.
- 📊 €1.25 billion in sales last fiscal year from those same people, now working as advisers.
Anyone can achieve the first figure. Almost nobody achieves the other two. They aren’t the same kind of number — one is a saving, the other is revenue — but the order of magnitude is the point: the business created by redeploying people is roughly a hundred times the saving.
What IKEA did
In 2021 it launched Billie, an AI assistant for routine queries: orders, opening hours, deliveries, returns. Today it can assist 74% of the customers who use it.
The usual move would have been to reduce headcount. IKEA asked a different question: what is in the part AI doesn’t resolve well?
It found a pattern. Customers were asking for help to design and plan their homes — decisions that need judgement, listening, and someone who knows what they’re talking about.
So it reskilled its agents as remote interior design advisers, trained in planning tools and in leading an advisory conversation. Today they are among its fastest-growing sales channels, growing 15-20% a year. They serve 10 million customers a year. Its internal customer happiness score rose from 60% to 89%.
AI paid for the project. Human intelligence turned it into a business.
It isn’t savings. It’s release
AI handles routine work well. But it doesn’t cover all the value the customer needs.
When you take agents out of chat and voice so AI can replace them, the need for advice doesn’t disappear — it goes unmet. In our own CX Gap 2026 research, only 29% of bot conversations are resolved without escalating to a person, and 46% of customers abandon self-service because the information is inaccurate or insufficient. That distance between what automation resolves and what the customer actually needed is the CX Gap.
We’ve seen that gap close with a different decision: not treating the human capacity AI frees up as a cost to eliminate, but as capacity available for what the customer asks for and the machine doesn’t provide — better advice.
AI doesn’t eliminate human intelligence. It frees it. The question is where you put it next.
What we see with video
Redeploying the agent as an adviser is the first step. The second is giving them the channel where advising pays off most: where they can see the customer, see the document, show the product. On the phone you describe; on video you show.
In a branch network where customers choose between in-person service and a remote adviser on video, with the same operations available, thousands of surveys show:
- ✅ Waiting time: from 20-25 minutes to 0-3 minutes at most
- ✅ Satisfaction: 9-10/10, on a par with in person and above any digital channel
- ✅ First contact resolution: 95%, the same as in person
- ✅ Capacity: +50% with no extra floor space and no extra staff
And the adviser on video is rated higher than in voice or automated service. The work is more complete — and it shows in the person doing it.
Advised Service
The lesson fits in one sentence: AI isn’t there to save on human intelligence. It’s there to put it where it creates the most value.
That value has a name: Advised Service — service in which the person doesn’t just answer, but guides a decision: what to buy, what to choose, how to solve it.
Video is what makes that redeployment pay off in full. Remotely, from the contact center. Or at the point of service itself — store, branch, clinic — with the customer present and the specialist one tap away.
AI resolves the routine. When the query needs judgement, it passes to a person with all the context, so the customer doesn’t start from zero. And that person advises from the channel that gives them the most capacity.
Two intelligences — AI × HI (Human Intelligence) — each where it is strongest.
💡 When your AI frees up your team, what capacity will you give them to advise?
Sources: Fortune (July 30, 2026); Ingka Group Newsroom (June 2023); Interactive Powers, The CX Gap 2026; anonymized Interactive Powers deployment data.
Interactive Powers - Streamline your business communications