Why AI Handling Routine Customer Interactions Is a Redeployment Story, Not a Layoff Story

Numa’s AI Receptionist and Smart Inbox already handle the routine, high-volume share of customer communication for 1,300+ dealerships, more than 1 billion calls, and the dealerships getting the most out of that shift aren’t the ones cutting headcount. They’re the ones redeploying their existing team toward the work AI can’t do. That distinction, between seeing AI as a reason to shrink a team and seeing it as a reason to redirect one, is exactly what separated two very different outcomes at two very different companies recently, and it’s the same choice facing every GM deciding what AI adoption actually means for their service department.

Ikea Automated Half Its Calls. It Didn’t Cut 8,500 Jobs.

When Ingka Group, Ikea’s largest retailer, rolled out an AI assistant that ended up handling 47% of customer calls, the math on paper pointed toward eliminating roughly 8,500 roles. Ingka didn’t take that path. It kept the people and looked at what customers were asking for that the AI couldn’t provide, discovering real demand for in-home interior design help. Those 8,500 employees were retrained into that role instead, a new revenue line that hit €1.3 billion in 2024 and is projected to reach 10% of total revenue by 2028.

Klarna took the other path. An AI agent handling 2.3 million conversations a month cut resolution time from 11 minutes to under 2 and was projected to replace the workload of 700 people, until the gaps showed up in exactly the conversations that needed judgment: sensitive billing situations the AI handled poorly enough that frustrated customers started airing complaints on social media and Trustpilot. Klarna ended up rehiring, telling CX Dive its new philosophy directly: “AI gives us speed. Talent gives us empathy.”

The dealership version of this choice is playing out right now, at a smaller scale, in every service department deciding what “AI is handling the routine calls now” actually means for the people who used to handle them.

The Data Says Klarna’s Path Is the More Common Mistake, Not Ikea’s

Klarna’s reversal wasn’t an outlier. Gartner has specifically predicted that half of all companies that cut customer service staff because of AI will be forced to rehire, often under different titles, by 2027. That’s not a general AI-adoption prediction. It’s specifically about customer service roles, the exact category a dealership’s BDC and service advisor positions fall into.

The pattern is already showing up broadly. A Robert Half survey of 2,000 U.S. hiring managers, reported by Fast Company, found 32% had eliminated a role due to AI-driven productivity gains and later rehired for that same position. Forty percent of those who rehired said the role needed institutional knowledge or context AI couldn’t replace, and 38% said AI required more human oversight than expected. Separately, Forrester Research found 55% of employers regretted their decision to cut staff because of AI. Robert Half’s Megan Slabinski summarized the pattern directly: companies are learning where AI works best alongside their employees, not in place of them.

What This Actually Means at a Dealership

None of this means AI isn’t handling real volume. It is, and the volume is substantial. Dealership service departments routinely miss 300 to 500 calls a week, most of them status checks and scheduling requests that never required a live advisor’s judgment in the first place. The question a GM has to answer isn’t whether AI can absorb that volume. It’s what happens to the time that volume used to consume once it’s gone.

That question has a genuine answer, and it isn’t “nothing.” Dealerships already running AI alongside their existing team, rather than in place of it, report freed advisor and BDC time going toward exactly the conversations Klarna discovered its AI couldn’t handle: a customer disputing a charge, a multi-vehicle trade negotiation, a repeat complaint that needs real problem-solving, a declined-service follow-up that could become a sold job with the right conversation. None of that requires new headcount. It requires the same people spending their day differently.

Where the Freed Time Actually Goes

Toward the customers already at risk of leaving. Equity mining and declined-service follow-up are two of the highest-value activities a BDC or service team can do, and both routinely go undone not because nobody wants to do them, but because there’s never time left after the routine call volume is handled. Freeing that time doesn’t eliminate the job. It redirects it toward the outreach that actually protects revenue.

Toward the complex conversations that determine CSI. Real-time sentiment monitoring exists to flag exactly the moment a routine interaction turns into something that needs a person, mirroring the same gap Klarna discovered the hard way. A dealership that redirects freed time toward those flagged conversations, rather than eliminating the seat that used to handle them, is applying the Ikea lesson directly: find what the technology can’t do, and put people there instead.

Toward relationship work that was always the highest-value use of an advisor’s time anyway. Service advisors are typically already scheduled close to capacity before a single call comes in, 12 to 15 repair orders a day with roughly 15 minutes at write-up and 5 to 10 minutes at delivery for each one. Removing status calls doesn’t create idle time. It restores time that was already overcommitted, time that goes back into the diagnosis conversations, the upsell conversations, and the walk-up customers who were previously competing with a ringing phone for the same advisor’s attention.

What Dealerships Doing This Already Look Like

The redeployment argument isn’t hypothetical at the dealerships already running this way. One GMC dealership put it plainly after its best customer-pay day in the store’s history:

“They’ve had their highest customer pay day ever and a huge increase in appointments booked. They have made no other changes and attribute this to Numa.”
— GMC dealership

A Honda service manager described the same pattern from the retention side, a store that redirected its existing advisors’ attention rather than restructuring around smaller headcount:

“The product does everything we were sold on, and we’ve had the same team since the beginning. Everyone is so involved and provides consistent communication and support.”
— Service Manager, Honda dealership

That mirrors what Numa CEO Tasso Roumeliotis has argued directly about AI adoption inside dealerships: resistance to AI is consistently a leadership problem before it’s a technology problem, and staff who fear a new system means their role is next need that question addressed directly, with a real answer about where their time is actually going, not vague reassurance.

The Numa POV: The Redeployment Question Is the Only One Worth Asking

Ikea and Klarna didn’t end up in different places because one company’s AI was better than the other’s. They ended up in different places because one asked what its people could build once the routine work was handled, and the other asked what could be cut. Numa was built around the same premise Ingka landed on: AI absorbs the volume that never needed a person, voice, text, workflow triggers, real-time sentiment monitoring, so the humans on a dealership’s team spend their day on the conversations that actually determine whether a customer stays. A GM who frames AI adoption as a redeployment question ends up with a team doing higher-value work. A GM who frames it as a headcount question is making the same bet Klarna made, and the data says that bet is wrong more often than it’s right.

Frequently Asked Questions

Does AI in a dealership’s service department actually lead to layoffs?

The data suggests the opposite is more common when AI is applied without a redeployment plan. Gartner has predicted that half of companies that cut customer service staff specifically because of AI will be forced to rehire by 2027, and a Robert Half survey found 32% of companies had already eliminated a role due to AI and later rehired for it. The pattern shows up because AI reliably handles routine volume but consistently underperforms on conversations requiring judgment, institutional knowledge, or empathy.

What happened when Klarna replaced customer service staff with AI?

Klarna’s AI agent handled 2.3 million conversations a month and was projected to replace roughly 700 employees’ workload, cutting resolution time from 11 minutes to under 2. Within a year, gaps in handling sensitive billing conversations led to public customer complaints, and Klarna began rehiring human staff, stating that AI provides speed while human talent provides empathy.

How did Ikea handle AI without cutting jobs?

Ingka Group, Ikea’s largest retailer, automated 47% of customer calls with an AI assistant that could have justified eliminating roughly 8,500 roles. Instead, it retrained those employees to provide in-home interior design services, a new revenue channel that reached €1.3 billion in 2024 and is projected to grow to 10% of total revenue by 2028.

If AI handles routine service calls, what should the freed-up staff time be used for?

The highest-value uses are typically declined-service and equity-mining follow-up that often goes undone due to lack of time, complex or frustrated customer conversations flagged by real-time sentiment monitoring, and the diagnosis and relationship conversations advisors are already scheduled close to capacity for before any calls are even added. None of this requires new hires; it requires redirecting time that was previously consumed by routine call volume.

Why do GMs and service advisors fear AI will eliminate their jobs?

The fear is reasonable given how AI adoption is often marketed, as a direct headcount replacement. Research consistently shows the more common real-world outcome is a redeployment of existing staff toward higher-judgment work, not elimination of the roles themselves, but that outcome depends on leadership explicitly communicating where freed-up time is going rather than leaving staff to assume the worst.

See how Numa frees your team’s time for the conversations that actually need them. Talk to Numa.