Will AI Replace Retail Jobs? Not on the Luxury Shop Floor

Earlier this year a small boutique opened in San Francisco with an unusual manager. Andon Market is run by an AI agent called Luna, installed by Andon Labs on a three-year lease. Luna's first decision, as Forbes reported in May, was to hire people: humans to handle the customer and everything that happens in person.

That is a better summary of AI in retail than most forecasts managed. For years one statistic has circulated through career sites and conference decks: 65 percent of retail jobs could be automated by 2026. It is now 2026, and the shop floor is still staffed.

The prediction has expired

US retail trade employed 15.39 million people in September 2019, according to the Bureau of Labor Statistics. In September 2026 it employed 15.38 million. Seven years, a pandemic, the arrival of generative AI, and the headcount of American retail has barely moved.

The wider economy tells the same story. Yale's Budget Lab, which tracks AI's effect on the labor market, found in May that there were no clear effects yet on overall US employment. Whatever AI is doing to retail, it is not doing it through mass layoffs.

Where the machines actually landed

That does not mean nothing has changed. It means the change is concentrated, and the new BLS projections for 2025 to 2035 show exactly where.

• Cashiers: down 200,600 jobs, the largest projected numeric decline of any occupation in the country. BLS attributes it to technology, chiefly self-checkout and online sales.
• Customer service representatives: down 141,800, as chat and voice tools absorb routine enquiries.
• Shipping, receiving and inventory clerks: down 62,800.
• First-line supervisors of retail sales workers: down 52,500.
• Retail salespersons: roughly four million jobs, down 14,000 over the decade. BLS rounds that to zero.

Read the list as a map. Automation is taking the transaction, the back room, the call centre and part of the supervisory layer. It is not taking the person who sells. A survey of more than 250 retail leaders by Retail Economics points the same way: the biggest potential for AI sits in HR, finance, compliance and reporting, not on the floor.

Luxury never sold the transaction

If that is true of retail at large, it is truer in luxury, which never had a transaction worth automating. Nobody self-checks out of a high jewelry salon. The sale is a conversation that may run across several visits, and the value sits in the relationship rather than the till. We covered how that is reshaping skills in our piece on AI clienteling; the employment data now confirms it at scale.

LVMH employed 43,193 people in the United States at the end of 2025, across 1,232 stores. Its third-quarter update this week named the US among its markets with good growth. McKinsey's May report on AI in luxury concludes that in-store AI works best behind the scenes, surfacing context for the adviser while leaving the moment of commitment to a human.

Kering's restructuring shows where the technology money is going instead. In March it created a group-wide Client center of excellence, covering product, pricing, distribution, planning and data, and put a former McKinsey and Amazon executive in charge as Chief Client Officer. The intelligence is being centralised at group level. The floor keeps its people.

What actually changes in the job

The jobs survive. Their content does not stay the same, and candidates who miss this will interview badly. Three shifts matter. First, memory moves into the system. An advisor used to win by remembering the client's daughter's wedding date; the CRM now remembers it for everyone. What remains scarce is judgment: knowing when not to call, which suggestion to ignore, how to read a client the data has misread. That is the core of client development now.

Second, the stock room is where automation lands hardest. Counting, receiving and replenishment are being absorbed by systems. The stock and operations roles that hold up are the ones built on exceptions: allocating scarce pieces, managing client reservations, keeping a flagship's back of house running at the pace the floor demands.

Third, management is thinning in the middle. When software handles scheduling and reporting, a store needs fewer people supervising and more people coaching. Store leadership in luxury is drifting toward the second job.

What we screen for now

Our interviews have shifted with the work. The questions that separate candidates today:

• Book evidence, not volume. How many clients came back, and why.
• Tool fluency without dependence. Can you use the CRM, and can you say when its suggestion was wrong?
• Discretion as a skill. Data makes it easier to overreach; the best advisors know what not to say.
• Coaching for managers. Can you develop an advisor, not just schedule one?

None of this has lowered the value of the role. Our US luxury retail salary guide still shows the premium sitting with advisors who own a client book, and the sales ladder remains the main route into management.

The shop floor is safer than the back office

The honest answer to whether AI will replace retail jobs is that it already has: at the checkout, in the call centre, in the stock room and in a slice of middle management. In luxury, those cuts make the advisor more central, not less. The industry that sells relationships is the last one that can automate them away. Current openings sit on the jobs board.

Rebuilding a team around clienteling and data? Talk to us about your search