KI-Werkzeuge für Vermögensverwalter: Eine veraltete Momentaufnahme, kein Urteil

AI tools for wealth managers have arrived quietly. Every wealth manager I speak to now uses at least one AI assistant. Almost nobody has written down which one, for what, and with which data. This is my attempt to put that on paper.

A Snapshot, Not a Verdict

Everything below reflects the position on 13 August 2026. That date matters more than usual. These products change monthly, prices move, and a model that felt slow in spring can feel quick by autumn.

So please read the stars as one practitioner’s impression after daily use, not as a benchmark, not as a test result, and certainly not as a recommendation. I have run no controlled comparison. I have simply worked with these tools and formed a view, and my view will age.

If you need a real assessment for your firm, you need your own trial, your own use cases and your own risk analysis. That is exactly what the regulator expects, as we will see.

The Five AI Tools for Wealth Managers, and What They Cost

These are the published consumer rates I could see on the day and exclude VAT, team plans and enterprise agreements. Check them yourself before you budget anything.

The Criteria

I picked six criteria that reflect how our industry actually uses AI tools for wealth managers: drafting, research, working through long documents, grasp of the Swiss regulatory context, the confidentiality posture, and price.

Two rows deserve a comment rather than a star.

As for price, the Chinese tools win on the sticker and that is genuinely useful for private study, language work and general reading. It tells you nothing about whether they belong anywhere near a client file.

On confidentiality, meanwhile, nobody gets five stars. Not one of these consumer products is a place for client-identifying data, and the three stars simply reflect that the American vendors publish clearer enterprise terms, data-residency options and retention controls than the free Chinese apps do.

What FINMA Actually Expects

This is where the conversation usually stops being fun and starts being useful.

Unter Guidance 08/2024, published on 18 December 2024, FINMA set out what it expects from supervised institutions that use artificial intelligence. It is short, it is readable, and it applies to independent wealth managers, not only to banks.

The themes are straightforward. Governance and clear responsibility, so somebody owns the risk. An inventory of AI applications with a risk classification, so you know what you are running. Data quality, because incomplete or outdated inputs produce confident nonsense. Testing and ongoing monitoring, including fallback arrangements. Documentation. Explainability, meaning you must understand well enough how an application works to judge its risks. Independent review by people with the skills to do it. And, where the tool comes from a vendor, the ordinary outsourcing and third-party rules apply.

Now read that list again with a free chat app in mind. An unlogged, uninventoried tool that any employee can open in a browser fails most of it before you have typed a word.

None of this makes AI forbidden. Our own view on AI in der unabhängigen Vermögensverwaltung has always been that the technology is useful and the governance is the work.

The Part I Care About Most: What You Type In

Please be careful what you paste into AI tools for wealth managers.

A client name, an account number, a passport scan, a portfolio statement, a structure chart, a draft PEP assessment, an internal memo about a difficult relationship. Every one of those leaves your control the moment it goes into a consumer chat window, and no setting you toggle afterwards brings it back.

Swiss banking secrecy, data protection law and Pflichten zur Bekämpfung von Geldwäsche do not pause because the interface looks like a messaging app. Nor do grenzüberschreitende Vorschriften, which is a point worth thinking through when the servers sit in another jurisdiction entirely.

The practical rule we use is simple. If you would not send it to an external consultant without an agreement in place, do not paste it into a chat window either. Anonymise, generalise, or do not ask.

Where AI Tools for Wealth Managers Genuinely Help

With that boundary respected, however, the honest answer is that they help a great deal.

Tightening a paragraph. Turning a rambling voice note into a clean agenda. Explaining a structured product to yourself before you explain it to somebody else. Drafting a first version of a policy you will then rewrite. Working through a public regulatory text. Preparing questions for a fee conversation. Translating a market commentary into a second language without losing the tone.

What they do not do is replace judgement, and they are still perfectly capable of inventing a citation, a figure or a rule with total confidence. That is not a bug you can configure away. It is why the human part of this business has not gone anywhere.

Putting AI Tools for Wealth Managers on a Footing

First, write down which tools your firm actually uses. You will be surprised. Then decide which two you support and pay for properly, on business terms rather than private subscriptions. After that, write one page on what may and may not go in. Train the team once, then again in six months, because habits form quickly. Keep the inventory current.

In short, that is not a heavy programme. It is roughly a morning, and it turns an uncontrolled habit into a documented one.

A Closing Caution

Therefore, treat the table above as a conversation starter with your own team, rather than as a shortlist. I have not audited these vendors, I have no commercial relationship with any of them, and I would not want anybody choosing a tool for a regulated business on the strength of a blog post.

Instead, test them yourself, on your own material, with data you would be comfortable seeing in a newspaper. Then decide.

And write the date on your assessment. In this field, an undated opinion is already wrong.