Journal / Kirill Naumov
Best Buy’s AI Sales Training: The Demo Is Not the Customer’s Decision
Best Buy trained store specialists to demonstrate AI PCs. What should retail sales training test when the product itself is hard to explain?
By Kirill Naumov ·
A customer walks into an electronics store and asks whether an AI PC will actually help with her work. A salesperson can show a dazzling demo in thirty seconds. The harder part is asking what she does all day, whether her files can be used with the feature, what works offline and whether a cheaper laptop would be enough. If training stops at the demo, the team learns to perform a launch, not to help someone buy well.
Best Buy said in June 2024 that more than 30,000 Blue Shirt and Geek Squad colleagues had been trained and certified on Copilot+ PCs. It planned demonstrations and training in hundreds of stores, with dedicated computing specialists in many locations. That is a company statement about preparation and scale, not evidence that every customer was well advised or that any particular training method increased sales.
The case is useful because AI products change faster than most retail training programmes. A feature can look similar on two devices while differing in processor requirements, privacy settings, account needs or language support. A neat script goes stale. So the question is not “How many product facts can a colleague repeat?” It is “Can they find and check the answer when a customer’s use case does not match the marketing slide?”
Teach the question before the demonstration
Start with three customer situations: a student comparing battery life and price; a freelancer who handles confidential client material; and a family buyer who mainly needs video calls and office software. The same AI demo may be compelling to one and irrelevant to another. Ask colleagues to explain the difference in plain language, including what they do not know.
For each feature, build a small proof card: the task it helps with, the devices and accounts required, where data goes, the likely limitations and the approved source for current details. Make the card easy to update. A good salesperson should be comfortable saying “let me check the current requirement” rather than inventing a confident answer.
Practise a decision, not a pitch
An AI role-play can simulate a customer who has seen a viral clip and expects the PC to do everything. Score the colleague on listening, the fit between need and product, factual accuracy, a fair comparison and a clear next step. Do not reward a longer pitch or more AI jargon. One strong answer may be “you do not need to replace your laptop yet”. That can be good service even when it is not the quickest sale.
Include a handoff: when should a specialist step in, and what should the first colleague pass along? If the customer has a technical compatibility or privacy question, the right outcome may be a verified answer from the specialist, not an improvisation at the shelf.
Measure the sale after the customer gets home
Training attendance and demo counts are easy to collect. Add harder measures: returns caused by expectation mismatch, setup-related contacts, customer understanding after a demonstration and whether staff correct an outdated claim. Compare similar product categories and periods before attributing changes to a new training programme. Ask colleagues which questions still leave them uncertain.
The wider lesson from Best Buy’s public example is that selling AI requires a team that can translate a feature into a useful decision. The most impressive demo in the store is worth less than an accurate answer to a customer’s ordinary question.
Explore the English AI for Retail workshop → · Discuss AI sales training
Primary source: Best Buy’s June 2024 account of specialist training. Training scale is company-reported; the scenarios and assessment criteria here are independent proposals.