Professional services
Trust and Fiduciary Advisor
Trust companies in Valais — what changes by 2030
7 min read · 60% of tasks automatable, 100% of the trade transformed
The trust company is the most exposed link in the canton's economic fabric. AI is not going to make it disappear: it will erode what it used to bill for and revalue what it used to give away, advice and trust.
The trade today
Valais counts several hundred trust companies, the large majority of which employ fewer than ten staff; the Valais section of TREUHAND|SUISSE (the Swiss trust and fiduciary association) numbers around 80 members, firms run by their owners, averaging six employees each¹. This fragmented, regional fabric handles the bookkeeping, taxation, and often a share of the administrative life of thousands of the canton's SMEs, self-employed people, communes, and families. Chapter 10 of the essay makes this the ground on which its economic thesis finds its fullest expression.
A firm's daily work spans a wide range:
- Bookkeeping: data entry, reconciliations, interim and annual closings
- Taxation: returns for individuals and companies, VAT, withholding tax, objections and representation
- Payroll and social insurance: statements, certificates, AHV/BVG/UVG filings
- Audit: limited statutory audits for SMEs subject to the requirement
- Advisory work: company formation and restructuring, business succession, retirement planning, real estate
- Representation: liaising with cantonal and federal tax authorities, banks, and insurance funds
- File management: filing, supporting documents, follow-ups, correspondence
A substantial share of revenue today rests on intermediate tasks billed by the hour: data entry, preparing closings, standard tax returns. That is exactly the layer the current transformation is attacking.
What AI is preparing
Entry-level bookkeeping. Automatic reading of documents, suggested account coding, bank reconciliations, preparation of closings: what used to take days of data entry shrinks to a validation workflow. The trend, already set in motion by bank digitization, is accelerating sharply. A firm that bills data entry by the hour watches that revenue shrink; one that has shifted to flat fees and advisory work absorbs the transition.
Standard filings and correspondence. Simple tax returns, routine objections, correspondence with the authorities: drafted by AI from the file, then reviewed and signed by an accountable professional. Liability remains attached to a named individual; that is what the client pays for.
Tax and legal research. Cantonal practice, case law, doctrine: what used to take hours now takes minutes, with the new obligation to verify every reference produced (the risk of hallucination is greatest exactly where precision carries liability).
Pre-drafted advisory work. Retirement simulations, business succession scenarios, comparisons of legal structures: AI prepares the costed variants, and the trust advisor chooses which ones deserve to be presented and takes them to the client. Mandates once reserved for large firms become accessible to a six-person practice; the competitive shift plays out here in both directions described in chapter 10, as erosion and as conquest.
Client data: the prerequisite
A trust company's files concentrate what an SME and a family hold most sensitive: wages, wealth, disputes, plans. Three requirements precede any deployment: nFADP compliance (hosting, subcontracting, and transfers all documented; a free consumer-grade tool is not a working environment); compliance with AMLA obligations for financial intermediation activities, with traceable processing; and a contractual duty of discretion explicitly extended to the tools themselves, so the client knows what is being processed, where, and by what. The practical sovereignty discussed in chapter 10 is decided precisely here: tools that comply with Swiss law, data that stays on Swiss territory, a chain of accountability under Swiss jurisdiction.
What rises in importance for judgment
Validation that carries liability. A signed closing, a filed tax return, an audit report: acts that bind the professional's responsibility, regardless of which tool prepared them. Reviewing a generated deliverable with the eye of the person who will sign it is a distinct skill from drafting, and it is becoming central.
Reading the client behind the file. Knowing that the joinery workshop's abnormal margin is explained by the owner's illness, that a certain self-employed couple is preparing to separate, that a particular figure a client declared deserves a question before it is passed on to the tax authorities: this knowledge appears in no structured dataset.
Contextual tax judgment. Between defensible optimization and the risk of a reassessment, between cantonal practice and the letter of the law, the line is drawn through experience with past files and knowledge of the authorities. AI cites; the trust advisor calibrates.
Succession advice. Valais business successions over the next ten years (a matter of owner demographics) will play out in conversations where wealth, family, and emotion intermingle. That is not a conversation you send a tool into.
Trust built over decades. Thirty years of closings for the same family of craftspeople constitute an asset no platform will ever match. Provided it is not left dormant: it is exactly what justifies the shift from hourly billing to an advisory retainer.
Who keeps the final say?
| AI proposes | The trust advisor judges | The firm bears responsibility for |
|---|---|---|
| An annual closing pre-drafted from bank flows and scanned documents | Whether the entries reflect the company's economic reality, whether provisions and depreciation express a strategy and not merely a rule | The fidelity of the accounts and responsibility toward the client, the bank, and the tax authorities |
| An optimized tax return with three costed variants | Which one is defensible before the cantonal authorities, given local practice and the taxpayer's history | The risk of a reassessment and the long-term relationship with the tax authorities |
| A summary of case law on a business succession matter | Whether the references are accurate, whether the case truly applies, and what the family is ready to hear | The advice given and its consequences for the estate and the family |
| An anomaly alert in a client's accounts (margin, cash flow, unusual movements) | Whether it reflects an error, a life event known to the firm, or a signal that calls for a conversation, or even legal obligations | The quality of the audit and the firm's obligations, notably under AMLA |
Composite illustration. A six-person firm is preparing the closing for a family joinery business. The tool flags a falling gross margin and proposes, with sound accounting logic, a provision and a cautiously worded cover letter for the bank. The partner who has followed this client for fifteen years knows the owner spent six months away from the workshop for health reasons and that his son has just taken over running it. She calls before she writes, learns about the spring orders, and turns the bank letter into a generational transition file with a two-year plan. The number told a story of decline; the file told a story of succession. (A fictional, composite situation, to be replaced by a real, anonymized case during the incarnation pass.)
What AI does not do
It does not protect the historic business model. A firm can deploy the best tools available and still see its revenue fall if its billing remains built on the hours those tools eliminate. Moving from hourly rates to a flat fee, an advisory subscription, or value-based mandates is a strategic decision that technology makes urgent without making it for the partners. Nor does it replace succession planning: recruiting and training the generation that will take over the firm remains the primary risk facing Valais trust companies, with or without AI.
Territorial anchoring
Pressure comes from two directions: global accounting platforms quietly absorbing simple mandates, and large firms from urban centers moving down into mid-sized files. The competitive shift makes a response possible: a well-equipped six-person firm, backed by its knowledge of the cantonal terrain, can hold complex mandates and deliver to a Geneva client work that an international firm used to provide. The regional advantage, discretion, proximity, long-standing trust, cantonal practice, becomes a commercial argument again, provided it is properly equipped. For mountain communes and SMEs, the presence of a competent local trust-company network is also a matter of viability infrastructure: it is often the only remaining local advisory service once the bank has closed its counter.
What decision-makers must do now
For a firm partner
Before the end of 2026, measure the share of revenue built on tasks that AI is absorbing (data entry, standard closings, simple filings), and decide on a three-year billing trajectory. Then set down the firm's rules (practice, tone, thresholds of caution) in a controlled instruction file, and train the senior partners first: they are the ones who know what a deliverable needs to be worth.
For TREUHAND|SUISSE, Valais section
Champion the "trust advisor" career track at the alpine campus (a pilot sector under PA-I1): case exercises on fictional data, shared rules on nFADP/AMLA compliance, and experience-sharing between firms. Pooling training prevents every small firm from paying its own learning curve alone.
For the Cantonal Tax Administration
Publish a clear position on AI-prepared filings and correspondence (accepted formats, traceability requirements, the signatory's responsibility): legal certainty for firms determines the healthy pace of their adoption, and the administration gains better-structured files in return.
¹ TREUHAND|SUISSE, Valais section: around 80 corporate and individual members, firms averaging six employees (section website, accessed 2026). The total number of active trust companies in the canton (members and non-members) runs into the hundreds, a figure drawn from the essay (ch. 10), still to be confirmed.
Jérôme Deshaie is the founder of MCVA Consulting SA, an agency specializing in the AI transformation of organizations in Valais, and the author of Bisse Cognitif.
The French version is authoritative.