Retail Salesperson

Retail Salesperson scores 43/100 on this site's structural exposure measure, where lower means more exposed (Medium confidence, two independent pillars). That is the moderate-risk band, not an extreme reading. Both the OpenAI "GPTs are GPTs" task ratings and the AIOE index land slightly on the exposed side of their reference populations.

The score describes tasks, not headcount: it cannot say whether AI will replace salespeople or only change how they work, only that a good share of what they do, such as answering product questions and recommending items, is the kind of work language models handle.

Real-world usage points the same way. In Microsoft's "Working with AI" study of Copilot conversations (arXiv 2507.07935), this occupation scored 0.299 on AI applicability, against an average of 0.159 across 785 occupations, which this site bands as High. The Anthropic Economic Index gives it an observed-exposure value of 0.322, higher than about 93% of the 756 occupations in that file. Both are directional signals about chat usage, not evidence that anyone's job has gone.

The headcount data show no collapse. BLS Employment Projections Table 1. 2 projects retail salesperson employment to move from 3,998,100 in 2025 to 3,984,100 in 2035, a change of -14,000 jobs or -0. 3%. The Occupational Outlook Handbook's combined Retail Sales Workers page (which also covers parts salespersons) projects 0%, little or no change, and about 550,600 openings a year, mostly from workers who transfer to other occupations or leave the labor force.

BLS gives a cause in that section's Employment paragraph: it expects online sales to keep growing, limiting growth in the number of physical stores, and says competition from e-commerce is expected to reduce demand for retail salespersons. It adds that they should still be needed, since many shoppers still visit stores and the job covers a versatile range of customer service tasks. BLS does not mention AI. The retail salesperson median wage was $17. 03 per hour in May 2025.

The industry mix inside the projection is more telling. Using the BLS national employment matrix for this occupation, our own reading of the figures is that selling jobs shrink in categories where shoppers often know what they want and buy online, and grow where advice is still part of the purchase. Electronics and appliance retailers fall 11. 3% (113,300 to 100,400 jobs), sporting goods retailers 8. 2%, and furniture and home furnishings retailers 7. 7%.

Health and personal care retailers rise 6. 0%, building material and supplies dealers 3. 1%, and automobile dealers 2. 8%. BLS names e-commerce as the reason for the occupation as a whole, but it does not explain the industry split, so tying it to how much advice a purchase needs is our reading, not a BLS finding.

What big retailers say they are doing with AI is mostly assistance. Home Depot announced on September 1, 2026 that its Magic Apron assistant now runs in all of its more than 2,000 U. S. stores, reachable through the app or a QR code on store signs. Total Retail's report describes the tool as complementing store associates rather than replacing them, and executive vice president Jordan Broggi said "our associates remain at the center of the shopping experience".

Lowe's executives said on their August 2025 earnings call, as reported by Customer Experience Dive, that scores for associate helpfulness and knowledge improved significantly after its Mylow Companion app reached associates, which lets a paint associate answer a garden-center question. These are company statements, not independent measurements.

Walmart gives the clearest statement on jobs, and it covers all of its associates, not only salespeople. A Walmart spokesperson told Fortune in September 2025: "Over the next three years we expect headcount to remain flat as roles evolve." CEO Doug McMillon said some jobs and tasks will be eliminated while others are added. That is a company forecast, and it says nothing specific about sales-floor roles.

We searched for the disconfirming case, a named retailer cutting sales-floor associates because of AI. We did not find one in the sources we could read. Recent retail cuts that surfaced in searches were corporate or support roles, and some coverage was aggregator-level, so we did not use it.

Why the verdict is still At risk: the rubric in RESEARCH-CHECKLIST. md applies mechanically. The score is below 60, and we found no licensing wall. O*NET describes a job with no credential requirement. Physical tasks such as preparing merchandise exist, but we found no evidence that they protect the selling task itself, and we did not read O*NET work-context ratings. Trust and accountability arguments cannot qualify a job on their own under the checklist.

Confidence is Low because the verdict rests on the structural score and the absence of a wall, while the real deployments we could verify assist associates and BLS projects a flat headcount.

What this means in practice, as our reading rather than a measured result: the exposed part of the job is generic product information and routine recommendations, which an app or an associate-facing AI tool can now deliver. The part that holds value is knowing a specific category deeply, helping with fitting, demonstration and installation questions in person, and closing a considered purchase.

Workers who build that depth or move toward roles with a team or a physical service component are better positioned than those who stay on generic floor coverage.