Skip to content
Nicolas GhrayebCOMMERCIAL & MARKETING LEADERSHIP

Personalisation at scale for luxury buyers: what AI should and should not do

In luxury real estate, personalisation is not a merge field or a longer segment list. It is the sense that someone paid attention. AI can help a team pay attention to more buyers, more accurately, as long as it is used for preparation and relevance, and not to imitate a relationship that a person has not built.

Luxury buyers notice being processed

Ask a senior sales consultant what a high-value buyer dislikes most, and the answer is rarely price. It is the feeling of being processed. A message with the right first name and the wrong assumption. A brochure sent for a unit type they dismissed weeks ago. Three people from the same team asking the same opening questions.

None of this is dramatic, and none of it is new. What is new is how easy it has become to produce personalised-looking communication in volume. That makes the real thing harder to recognise, and more valuable when it is there.

I think this is the right starting point for any conversation about AI and luxury buyers. The question is not how much we can personalise. It is whether the buyer would say, afterwards, that we understood what they were looking for.

Personal does not mean customised

Marketing teams often treat personalisation as customisation: the right name, the right image, a version of the email for each segment. That has its place, but a luxury buyer is rarely persuaded by it.

What they respond to, in my experience, is relevance, discretion and timing. Relevance means we remember what they told us: they want a quiet building, they are buying for the family, they care about handover before a certain date. Discretion means we do not display everything we know. Timing means we get in touch when we have something worth saying, and stay quiet when we do not.

None of those three is a technology feature. They are habits of a good sales team. AI can support them, but only if the team already agrees that they matter.

Start with what the buyer has actually told you

The most useful personalisation data is not a clever inference. It is what the buyer said, asked or chose to look at, held in a form the whole team can use.

In my work on CRM-driven pipelines, the gap I kept seeing was not a lack of data. It was that the useful part sat in call notes, WhatsApp threads and one consultant's memory. When that consultant was away, the buyer had to start again.

This is where AI can be genuinely helpful. It can read notes and conversation records and turn them into a short, structured picture: what the buyer is looking for, which projects and unit types they have asked about, what they have said about timing, what is still unanswered. The consultant checks it, corrects it and adds what only they know.

The principle I would keep: use what the buyer has knowingly shared with us. Be careful with anything inferred from behaviour that the buyer would be surprised to learn we were tracking.

The enquirer is not always the buyer

One of the more practical points about luxury purchases in the region is that the person who contacts us may not be the person who decides. A family member, an adviser or a broker may be shaping the shortlist. The end user and the investor may be different people, or the same person with two different sets of questions.

Earlier in my career, I worked with international agency networks and commercial sales leadership on high-net-worth customer acquisition across European, American and GCC feeder markets. The lesson I took away is that those markets do not ask the same questions or expect the same style of conversation, and that treating them as one audience with translated copy rarely worked.

The same applies to a Dubai luxury launch today. Buyers come from many places, and language, tone and even the order of information matter. AI can help a team prepare for that: drafting material in Arabic or English from approved content, suggesting which parts of a project story are likely to matter to a particular type of buyer, and flagging when we do not know who the decision-maker is. A person still decides what to say, and checks that the translation sounds like a native speaker wrote it.

Where AI helps most

If I were choosing where to apply AI first for luxury buyers, I would pick the places where it supports a person's attention rather than substituting for it.

Preparation. A concise buyer brief before a call or a private viewing: history, stated priorities, open questions, who else is involved.

Matching the right material. Choosing which floor plan, view, amenity or service detail to send from the approved library, instead of sending the whole brochure.

Drafting in the buyer's language. A first draft of a follow-up that refers to what was actually discussed, written for the consultant to edit and send.

Pacing. A reminder that a buyer asked for time until after a trip or a family decision, so nobody contacts them early. Personalisation includes knowing when not to write.

Learning across conversations. Themes that show what buyers keep asking about, so the brochure, the sales gallery and the broker briefing answer those questions directly.

In each case the output is a better-prepared person, not an automated message.

Discretion is part of the product

Luxury buyers pay for privacy as much as for square metres. That should shape how we handle their information.

Before AI touches buyer data, I would want clear answers to a few plain questions. What information is being used? Did the buyer share it for this purpose? Where is it processed, and who can see the output? How long is it kept? These are matters for legal and IT teams, and they need to be settled before a pilot, not after it. I would also check what the applicable data protection rules require rather than assume.

There is a quieter aspect too. A consultant who says "I understand you are looking at the second floor and want to avoid the road side" demonstrates attention. A system that quotes back a buyer's browsing history, or mentions a detail they never gave us, demonstrates surveillance. Knowing something and showing it are different decisions, and the second should be made by a person with judgement about the buyer.

The limit of scale

There is an obvious tension in the phrase "personalisation at scale" when the product is luxury. Part of what a buyer is paying for is that they are not one of many.

My answer is to use scale behind the scenes and keep the front of the conversation human. Scale belongs in preparation, organisation and consistency: every buyer gets a well-briefed consultant, accurate information and a timely reply. It does not belong in the tone of the relationship. Fewer, better messages will serve a luxury buyer better than a perfectly segmented stream of them.

I would also be cautious about letting a system suggest anything that sounds like a commitment: pricing flexibility, availability, handover, returns or service promises. These should come from approved documents and a named person. A confident, personalised mistake is worse than a generic correct answer.

How I would start

If I were advising a luxury developer's commercial team, I would keep the first step modest.

First, agree what personal means for the brand. Write down, in plain language, what a buyer should feel after contact with the team. This becomes the test for every use of AI.

Second, fix the record. Decide where buyer information lives, who maintains it and what a consultant is expected to capture after each conversation. A summary tool cannot be better than the notes it reads.

Third, start with preparation, not outreach. Give consultants a brief before each conversation and see whether the conversations improve. Move to drafted follow-ups only when the team trusts the briefs.

Fourth, put consultants in the design from day one. They know which suggestions feel right and which feel intrusive.

Fifth, review the output as a buyer would. Read a sample of messages and ask the same question each time: could this have gone to anyone, and would the buyer be comfortable with how we knew what we wrote?

What the buyer should notice

If this is done well, the buyer should not notice AI at all. They should notice that the consultant remembered what mattered, that the information was accurate, that nobody asked the same question twice, and that they were never pushed.

That is not a new standard for luxury. It is the old one, applied more consistently by a team with better preparation.

Key strategic takeaways

  • Treat personalisation as relevance, discretion and timing, not as a merge field or a segment list.
  • Use what buyers have knowingly told you, held in a record the whole team can use.
  • Do not assume the enquirer is the decision-maker, and prepare for different buyer markets and languages.
  • Apply AI first to preparation, material matching, drafting for consultants and pacing, not to unattended outreach.
  • Treat discretion as part of the product: know why data is used, and decide deliberately what to show a buyer you know.
  • Use scale behind the scenes and keep the relationship human. Fewer, better messages beat volume.
  • Keep promises, pricing and commitments with people and approved documents.

Sources & references

  1. The value of getting personalization right—or wrong—is multiplyingPublished: 2021-11-12 · Accessed: 2026-11-10
  2. Customer Data: Designing for Transparency and TrustPublished: 2015-05 · Accessed: 2026-11-10