A common assumption in wealth management circles holds that artificial intelligence will compress advisory margins by automating research, client communication, and routine analytical work. The assumption is partially correct. Generative AI has already substituted for substantial portions of equity research, document review, basic legal drafting, and competitive analysis. Wealth advisors charging fees for these information-based services face genuine margin pressure as the underlying labour becomes near-free. The assumption is also incomplete. The inverse implication, less discussed in industry analysis, is that as routine advisory functions become near-free, the value of what AI cannot do increases.

What does AI not do? AI does not walk into a Geneva office on the principal’s behalf to negotiate a delicate divorce settlement. AI does not read the room at a Monaco yacht show to detect that a competing brokerage is offering a Russian principal a structured transaction the family advisor needs to counter. AI does not anticipate a family rupture before it surfaces by reading the texture of a Christmas dinner conversation between two siblings. AI does not defend the principal’s interests in conversations the principal cannot personally attend. These functions resist automation in ways information-based functions do not. They depend on human presence, judgment, and reputation in specific physical and social contexts.

The principals who matter most in UHNWI advisory are not paying for information. They are paying for someone to walk into a Geneva office on their behalf, to read the room at a Monaco yacht show, to anticipate a family rupture before it surfaces, and to defend their interests in conversations they cannot personally attend. These functions resist automation in ways that information-based functions do not.

Brooke Harrington’s ethnographic work on wealth managers captures the substance of what UHNWI advisors actually do. Harrington spent years embedded with private client lawyers, accountants, and family office directors, documenting the working architecture of cross-jurisdictional wealth management (Harrington, 2016). The work she documents is overwhelmingly relational and discretionary, not informational and analytical. The advisor is the trusted node in a network of family relationships, jurisdictional regulators, financial counterparties, and operational service providers. The trust is built through years of physical presence and demonstrated discretion. It cannot be substituted by an AI system that processes information faster than a human can.

Caroline Knowles’ observations of London’s super-rich reinforce the analysis from a different angle. Knowles documents how the wealthy operate through dense networks of advisors who function as gatekeepers, problem-solvers, and identity-managers (Knowles, 2022). The advisor’s value is partly informational but primarily positional: the advisor occupies a specific node in the principal’s network, with access and authority that cannot be replicated by a generic information system. The Aman Members concierge cannot substitute for the family’s private client lawyer of thirty years. The Bloomberg terminal cannot substitute for the senior wealth manager who attended the principal’s father’s funeral and remembers which siblings did not speak.

What this means structurally is a bifurcation in wealth advisory pricing. Information-based advisory work (research, analysis, document drafting, routine compliance) is becoming a commodity priced near the underlying cost of AI compute. Relational and discretionary advisory work (cross-jurisdictional coordination, family conflict management, regulatory positioning, transaction negotiation in high-trust contexts) is increasingly priced at premium levels reflecting its non-substitutability. The advisor who provides only information-based work faces margin compression. The advisor who provides relational and discretionary work commands premium fees that may be higher in 2030 than they were in 2020.

The advisors who understand this shift are repositioning their offers around presence rather than analysis. Their pricing reflects not the time spent on the analytical work but the value of the relational positioning. The senior partner at a multi-family office charging 50 basis points on the family’s investable wealth is not pricing the analytical labour. The fee reflects the value of the partner’s network access, the trust accumulated over decades, and the discretion the partner brings to family conflicts and cross-jurisdictional structuring. AI does not threaten this pricing. AI threatens the analytical labour at the periphery of the partner’s work, which is increasingly delegated to automated systems anyway.

There is a specific category of advisor that faces acute threat: the mid-tier wealth manager whose value proposition has been primarily informational. The advisor who builds client decks, runs portfolio analytics, drafts client communication, and provides research summaries occupies the precise zone AI is automating. Unless this advisor repositions toward the relational and discretionary functions, the role compresses substantially over the next five years.

The advisory firms positioned for the AI transition are those investing in their senior partners’ ability to operate in the relational and discretionary zones while delegating the analytical labour to AI infrastructure. The firms losing the transition are those investing in analytical headcount that AI will increasingly substitute. Goldman Sachs Private Wealth Management, JPMorgan Private Bank, and the major multi-family offices have already restructured their senior partner roles around relational and discretionary work, with AI handling research and analytics. Mid-tier wealth managers that have not made this restructuring are facing margin compression.

For luxury real estate brokerages, the same dynamic applies. The information-based functions (property search, basic market analysis, transaction coordination) are becoming AI-mediated. The relational and discretionary functions (introducing buyers to off-market properties, negotiating sensitive transactions, managing cross-jurisdictional structuring) remain premium and non-substitutable. The senior broker at Compass, Christie’s, or Knight Frank who occupies a trusted position in the UHNWI principal’s advisory network commands compensation that AI does not threaten. The transactional broker whose value is primarily informational faces structural margin compression.

Information is becoming free. Presence is becoming priceless. The brokerages and advisors who understand this will reposition their offerings, their pricing, and their senior partner roles around the function. The ones still selling information-based advisory at premium pricing will find the pricing unsustainable as AI compresses the underlying labour cost. The advisors who survive the AI transition will be the ones who understand the function and command premium pricing accordingly.

References

• Atkinson, R., Burrows, R., and Rhodes, D. (2016) ’Capital City? London’s Housing Markets and the Super-Rich’, in I. Hay and J.V. Beaverstock (eds.) Handbook on Wealth and the Super-Rich. Cheltenham: Edward Elgar Publishing, pp. 225-243.

• Harrington, B. (2016) Capital Without Borders: Wealth Managers and the One Percent. Cambridge, MA: Harvard University Press.

• Hay, I. and Beaverstock, J.V. (eds.) (2016) Handbook on Wealth and the Super-Rich. Cheltenham: Edward Elgar Publishing. ISBN 978-1-78347-403-5.

• Knight Frank (2026) The Wealth Report 2026. London: Knight Frank Research.

• Knowles, C. (2022) Serious Money: Walking Plutocratic London. London: Allen Lane (Penguin).

• Paris, C. (2016) ’The Residential Spaces of the Super-Rich’, in I. Hay and J.V. Beaverstock (eds.) Handbook on Wealth and the Super-Rich. Cheltenham: Edward Elgar Publishing, Chapter 12, pp. 244-263.