Personal shopping training should teach where AI helps and where human judgement decides. AI makes large assortments and long client histories easier to navigate, so a personal shopper arrives prepared. Human judgement decides the selection. It edits many possibilities to the few that deserve attention, and it reads a brief that contradicts itself.
What should AI do in luxury personal shopping?
AI should do the work that scale makes hard for a personal shopper. That means sorting a large assortment, surfacing items consistent with a client’s history, and gathering the relevant facts before an appointment. Luxury assortments are complex and client histories long; memory alone cannot connect a client with something relevant.
That help has a limit. Identifying a possibility is not the same as making a recommendation. A system can say what resembles the client’s past; it cannot say whether the client wants more of it. Prediction is not understanding: only a conversation separates the client who wants another quiet piece from the one who has moved on. We set out that argument, with how a brand governs client data, in AI in Luxury Retail: Personalisation at Scale Still Needs a Human. Technology should leave the personal shopper, whom we call the advisor from here on, better informed and no less questioning.
Editing is the luxury skill
Editing, choosing the few pieces that deserve attention and saying why, is the skill that separates a personal shopper from a search result. Personalisation is usually associated with more choice. Expert personal shopping does the opposite. A client facing a large assortment does not need another long list; they need someone able to select for them.
Editing reduces complexity and shows confidence. The advisor is prepared to exercise judgement instead of handing the whole decision back to the client. That judgement stays flexible: if the client’s reaction shows the brief was wrong, the selection changes and is not defended. Curation of this kind is where the advisor’s value is most visible. It draws on product fluency, knowledge of the person, and the willingness to leave good things out.
Reading a brief that contradicts itself
Clients often ask for things that pull against each other. Distinctive but discreet. Familiar but different. Appropriate without feeling predictable. A client may ask for one item while describing a need that points elsewhere, or arrive certain of a product and discover in conversation that another direction works better.
These requirements do not reduce cleanly to product attributes, and a system resolves the contradiction by ignoring half of it. The advisor works with the tension instead: asking what “discreet” means to this client, watching the reaction to the first piece, and refining the brief as the appointment develops. Judgement also matters when the obvious recommendation is wrong. The item most consistent with past purchases may be the one the client has moved on from, and only recent conversation shows it.
Science prepares the appointment, Art decides it
In our methodology, The Art & Science of Clienteling™, the Science is the commercial discipline: client intelligence, preparation, and knowing which history matters today. This is where AI fits. It makes the relevant facts visible before the appointment so that preparation time goes on thinking instead of searching. What does the client already own, what has changed, and which requests remain open?
The Art is the human craft: discovery, presence, storytelling, judgement and the confidence to make a personal recommendation. This is where the selection is decided. We ask advisors to arrive with informed curiosity rather than an appointment designed in advance by a system. Prepared this way, the advisor uses what they know quietly, without reciting it, and can introduce something unexpected because they understand the client well enough to explain why.
What happens when AI and the advisor disagree?
The advisor should find out why before choosing a side. An AI recommendation may reveal a pattern the advisor has missed. The advisor may hold recent context that changes what the pattern means. Neither source is infallible, so the useful question is which information drives each view and whether it still reflects the client.
Ranked recommendations can create an impression of accuracy they do not have, and a confident list invites confident acceptance. Advisors should treat a ranking as a prompt for a question rather than an answer. Written outreach after the appointment deserves its own check before sending; we describe that in our page on generative AI in client outreach.
What this means for your team
Personal shopping teams get the most from AI when leaders decide first where human judgement changes the appointment. Find where advisors waste time searching for information, where better preparation would change the appointment, and where the edit visibly changes the outcome. Give the first to systems and protect the last. The luxury teams we train often find that better tools raise the visibility of advisor skill, because discovery, editing and storytelling become the whole job. Our method rehearses editing under a contradictory brief. This week, ask each advisor to take one upcoming appointment, prepare with whatever tools they have, and present a few pieces with a reason for each choice and each omission.
Book a Discovery Call or Request a Capability Assessment
Key takeaways
- AI narrows possibilities; a recommendation is not valuable just because it is technically personalised.
- Editing, the confidence to choose a few pieces and stand behind them, is the luxury skill.
- Human judgement matters most when the brief contradicts itself or the client is uncertain.
- Historical behaviour helps an advisor prepare; it does not say what the client wants now.
Frequently asked questions
How much of the appointment should be prepared in advance?
The facts, and a first edit. Before the appointment the advisor should know what the client owns, what has changed since the last visit and which requests are open, and should have a short selection ready. The selection itself stays open until the client has reacted to the first piece.
How should an advisor use AI suggestions before an appointment?
As a starting point for questions. Review what the system surfaced, ask why each item appears, and compare it with what the client said most recently. Keep what still fits, drop what only fits the past, and arrive with a short edit and an open mind.
Does the client need to know AI was involved?
Follow the brand’s policy on disclosure. Whatever it says, the recommendation the client hears should come from the advisor, in the advisor’s words, with a reason that reflects the relationship. Clients feel the difference between a system suggestion and a choice made for them.
Who resists an edited selection?
Some clients, who want to see everything before they choose, and some advisors, who fear leaving out the piece the client would have bought. Show the edit first and offer the wider range on request. Coach advisors to explain each omission; an explained omission reads as expertise, an unexplained one as a gap.