Why has AI Visibility become urgent for asset managers, family offices and wealth managers?

Because investors have changed where their research starts: 58% already use AI to research investments (Swissquote) and AI-driven search grew 4.2x in 12 months (Evolve Media), while 91% of asset manager content is still locked inside PDFs (IDC Innovators), invisible to the crawlers. Whoever structures their content now takes the answers about their niche before AI systems settle on a reference set that leaves them out.

What changes at each type of firm?

The three profiles compete for different mandates and therefore compete for different questions inside the AI systems. The most common mistake is treating AI Visibility as a single problem of "showing up more": what decides the outcome is showing up in the right questions.

Source: ReBo target-question mapping, 2026.

Asset manager: the thesis is the asset

Every month, an asset manager produces exactly the raw material AI systems are looking for: dense analysis, with numbers, primary sources and an explicit view. The problem is that this analysis usually ships as a single document covering several subjects. One edition that discusses rates, currencies and equities partly answers three different questions; three thematic pages each answer one of them in full, and it is the full answer that gets cited.

Family office: authority without exposure

The natural objection here is discretion, and it is a fair one. But discretion applies to the families you serve, not to the firm's expertise. Multi-family offices compete for mandates, and the questions that precede a mandate are conceptual: family governance, holding structures, succession, allocation across generations. Answering those questions in public builds trust before the first conversation, without naming a single client.

Wealth manager: the honest comparison

The questions asked by someone shopping for a wealth manager are comparative by nature: compensation models, conflicts of interest, flat fee versus trailer commissions, when it is worth paying for advice. Firms that publish that comparison honestly, including the cases where their own model is not the best fit, tend to get cited precisely because the text has the structure of an analysis rather than an advertisement.

What questions are your clients already asking AI?

The starting point of the work is not picking topics, it is surfacing the real questions. In the diagnostic, ReBo defines a universe of 30 to 80 prompts that clients, allocators, journalists and consultants actually type, and in which the firm ought to appear. They fall into four families, with very different intents.

Source: ReBo target-prompt methodology, 2026.

The brand family is usually the most revealing part of the diagnostic, and the most uncomfortable. When the AI finds no structured material about the firm, it does not say "I don't know": it answers with whatever it can find, which may be an outdated filing, an old news story or a description of a competitor with a similar name. Absence does not produce silence, it produces inaccuracy.

Why is financial services the furthest behind, and the biggest opportunity?

The knowledge exists in abundance — monthly content, reports, investment theses, proprietary studies — but it sits in the wrong formats: PDFs and websites that depend on JavaScript, the two modes of invisibility explained in why AI can't read PDFs. The result is an entire industry stuck at Level 1 of the AI Visibility framework.

There is a cultural reason behind it. In this industry, content has always served a relationship function: it is sent to people who are already clients, to maintain the bond and justify decisions. It was never built to be found, because acquisition ran through a different channel, usually outbound. The PDF is that logic made concrete: a closed document, addressed to a list.

The inversion is what creates the opportunity. When an investor asks an AI and gets three to five sources back, every answer is a shop window with very few slots. If almost the whole industry is invisible, the contest for those slots is effectively empty, and the cost of taking them today is a fraction of what it will be once the map of references has settled.

AI systems cite the best answer to the question, not the biggest brand. A smaller firm with well-structured content gets cited ahead of a large one with everything locked inside PDFs.

Where do social platforms fit in?

The principle that organizes everything here is simple: AI cites text, not aesthetics. On any platform, what becomes a source is real, structured text carrying a number and its origin, not an image and not a bare link. Social amplifies reach and generates the external authority signals; the definitive citable asset is still the page on your own domain, which everything should point back to.

The six rules that apply to every platform

• Text in the body, not just a link. Write the thesis and the numbers inside the post; bare links are barely read.
• One citable idea per unit. Every post, slide or tweet should answer something on its own.
• Number, source and date. "May IPCA-15, Brazil's preview inflation index: 0.62% (IBGE, the national statistics agency)" is worth more than "inflation went up".
• Named author. The portfolio manager's name and title reinforce verifiable authority.
• Always point to the full article. Social is the shop window, the website is the source.
• Thematic consistency. Publishing regularly on the same themes builds domain authority.

Source: ReBo analysis of platforms and citation factors, June 2026.

The practical implication is uncomfortable for anyone who invests in design: a beautiful carousel with the text baked into the image has human value and zero value for AI. It is the caption that carries the citable content. The same goes for a screenshot of a chart on X and for a video with no description.

The subject is regulated. How do you publish without taking on risk?

By publishing analysis and education, not recommendations. That is the line separating content that builds authority from content that creates regulatory exposure, and it sits in a more comfortable place than most firms assume, because it is precisely the analytical material, not the recommendatory kind, that AI systems cite.

Source: ReBo editorial guidelines for regulated institutions, 2026.

In practice this translates into workflow, not censorship: every piece is reviewed by the institution before it goes live, within a window agreed at the start of the project. The design exists to accommodate compliance review without stalling the publishing cadence. The step-by-step of the transformation is in how to turn the monthly letter into AI-friendly content.

What should you measure to know it is working?

AI Visibility is only worth the investment if it is verifiable, and the central metric is the citation rate: the percentage of target questions in which the firm shows up in the answer. It is measured against a baseline established before anything is published, over the same set of prompts, on a fixed cadence.

Source: ReBo metrics dashboard, 2026.

Two cautions keep the reading honest. First: citation is not deterministic, so a single answer measures nothing, which is why the test runs across dozens of questions over several rounds. Second: nobody controls what a model says, and anyone promising a guaranteed placement is selling something they cannot deliver. What can be controlled are the variables that move citation, and what gets measured is their effect, on the same ruler, month after month.

Where should you start?

With the diagnostic: finding out which ChatGPT, Perplexity, Gemini, Claude and Copilot answers your firm shows up in today, and which ones competitors occupy in your place. It defines the target prompts, measures the baseline and grades the firm against the framework, and it is included in the first conversation with ReBo.

From there, the path runs through the technical foundation and the release of the existing archive, usually the fastest return in the project because the material has already been produced and already been approved. Then comes the recurring cycle of transformation and distribution. The detail of each stage is in what ReBo delivers, and the conceptual rationale in the ReBo method.