Real estate broker

The Valais property market resembles no other in Switzerland: the federal act on second homes caps them at 20% of the housing stock in every municipality, while housing shortages for residents settle into the valleys. AI will not sell properties in the broker's place: it will produce in a few minutes the listings, files and valuations that used to fill entire weeks, and shift the broker's value toward what no model knows, the micro-market of a given municipality and the life story behind every transaction.

The profession today

The Valais real estate broker operates in a market split into segments governed by different rules: primary residences under pressure in towns and valley floors, second homes rationed under the Lex Weber in tourist municipalities, properties covered by grandfathered rights that circulate between the two, buyers domiciled abroad subject to their own legislation. Selling two comparable chalets in two neighbouring municipalities can mean two entirely distinct markets. The job demands holding together law, finance and local knowledge.

A broker's week splits between:

• Winning mandates: networking, referrals, preliminary valuations, local presence
• Valuation: property analysis, comparable transactions, regulatory constraints, report to the seller
• Marketing: photos, listings, brochures, portals, often in several languages
• Viewings and qualification: organising visits, accompanying prospects, sorting the merely curious from serious buyers
• Negotiation: between the seller's expectations and the reality of offers received
• Sale file: documents, easements, condominium regulations, coordination with the notary and the bank
• Administrative follow-up: correspondence, chasing, filing, agency bookkeeping

What AI is preparing

Multilingual marketing. Listing, brochure and sale file in French, German and English, generated from the file's documents and proofread before publication. The complete file (description, location, relevant excerpts from the condominium regulations, points requiring attention) comes together in hours rather than days. For a small agency, matching the presentation standards of the big networks becomes a matter of method rather than headcount.

Assisted valuation. Valuation models have existed in banks for years; generative AI puts them within reach of any agency and drafts the accompanying report. The limit is structural: these models learn from past transactions and measurable criteria, whereas in the mountains two properties that look alike on paper can be separated by a shaded slope, a cleared or uncleared access road, a neighbourhood. A generated valuation is only as good as the micro-market knowledge that corrects it. Without that correction, it is fast and wrong.

Virtual viewings and buyer screening. A digital twin of the property, remote viewings for prospects living hundreds of kilometres away, pre-qualification before any site visit: documented financing, a genuine residential project, compatibility with the property's status under the Lex Weber. Fewer viewings, better ones.

File back-office. Correspondence, chasing, gathering documents, preparing files for the notary: prepared automatically, validated before being sent. Hours given back to fieldwork and the phone.

Buyer and seller data: the prerequisite

A brokerage file concentrates sensitive data: buyers' income and financing, family situations behind sales (divorce, inheritance, moving into care), the history of offers received. All of this falls under the Federal Act on Data Protection, in force since 1 September 2023. Three requirements before any rollout: tools under contract with controlled hosting, ideally Swiss (copying a financing file into a free consumer-grade tool amounts to handing it to an uncontracted third party); defined retention periods, particularly for files of active buyers, an agency's most valuable asset and its greatest liability; and transparency with all parties about what is kept and what is deleted.

What rises in judgment

Micro-market knowledge. Knowing what has actually sold in the village, at what price and why, and everything the land register stays silent on: the arrangement between neighbours, the municipality's project for the parcel next door, a building's reputation. This knowledge cannot be scraped. It is built over years of presence on the ground.

Regulatory literacy. Explaining to a seller what their municipality's second-home quota means for their property, to a buyer what a home designated as a primary residence can never become, to heirs what the grandfathered right attached to their chalet is worth: the broker translates federal legislation into individual decisions. AI documents the texts; applying them to this property, in this municipality, remains interpretive work for which the broker carries responsibility.

Advising on life trajectories. Behind a transaction there is rarely a mere transfer of assets: a family of skilled newcomers choosing a valley, the population that chapter 12 describes as the one worth retaining; heirs scattered across three cantons parting with the family property; an elderly couple selling to move closer to care. The broker able to accompany these trajectories, listening, allowing time, saying when it is too early to sell, practises a different trade from the mere listings distributor. That is the one who stays.

Market ethics. In a tight market, overselling is easy and profitable in the short term. Chapter 13 makes housing the tipping point of Valais demographic policy: if residents can no longer find housing, the canton's attractiveness turns against itself. The broker sits at the front row of this tension. Giving an honest price, refusing an overvalued mandate, flagging that a property belongs to the resident market rather than tourist demand: these everyday choices carry public weight.

Negotiation. Two parties, money, often emotion. Bringing a seller to hear the market and a buyer to formulate a serious offer remains an art of relationship, one that the acceleration of paperwork makes even more central.

Who keeps the final word?

Composite illustration. An agency takes on the mandate for an estate: three heirs scattered across the country, a chalet in a municipality where the second-home quota has been reached. The tool produces the valuation, the sale file and listings in three languages within two days, work that used to take two weeks. The valuation, calibrated on comparable primary-residence sales, misses the essential point: built before the Lex Weber came into force, the chalet carries a grandfathered right and can be sold as a second home, in a rationed segment where demand outstrips supply. The broker documents the status, revises the valuation upward and explains it to the heirs, documents in hand. The sale closes in the segment the model had never seen. (Fictional, composite situation; to be replaced by a real case during the embodiment pass.)

Job profile 2030

The first new competency is supervising assisted valuation: understanding what the models know and do not know, documenting every correction made (slope, access, regulatory status), building over the course of files the market memory that makes up an agency's value. A broker who signs a generated valuation without correcting it takes responsibility for a figure they did not build themselves.

The second is governance of client data: ensuring that buyer files, financing dossiers and sales histories comply with the Data Protection Act, that tools are under contract, and that traceability holds up in the event of a dispute. In a small agency with no in-house lawyer, this vigilance rests on the broker alone.

The third is advising on residential trajectories: accompanying arrivals, inheritances and departures with an overall view (regulatory, financial, familial) that the automation of paperwork frees up the time to exercise. This part of the trade already existed among the best; it becomes the core of the offering.

Territorial anchoring

Housing is, per chapter 13, the tipping point: a canton can succeed in its shift toward AI, attract skilled workers, and still lose if its residents can no longer find housing. The broker works precisely on that ridgeline. Every mandate steered toward the resident market, every price held rather than pushed, contributes, at the scale of a single transaction, to keeping that tension governable.

AI changes the economics of the profession in the same movement: on documentary production (listings, files, reports), the gains are on the order of the factor of four to five the author has observed for this type of task, which will lower the cost of entry into the profession. What becomes scarce is the other side of the trade: knowledge of the municipalities, case-by-case reading of the Lex Weber, trust built over years. Tourism, which accounts for a seventh of the canton's GDP and whose transformation chapter 8 describes, rests on ground negotiated transaction by transaction; the quality of those who conduct these transactions is a matter of public interest.

What decision-makers must do now

For an independent broker or a small agency

Measure the time actually spent on listings, files and valuation reports, then equip these three chains first, with systematic validation before any file is sent. In parallel, bring the client and buyer file into compliance with the Data Protection Act (contracts, hosting, retention periods): this is the asset that will gain value as documentary production becomes commonplace.

For the Valais chamber of real estate

Establish an industry standard on assisted valuation: what a generated report must disclose (model used, comparables selected, human corrections made), so the profession keeps control of its credibility. Negotiate collective terms for compliant tools, and champion a real estate track within the expansion of the Alpine Campus (PA-I1), alongside the existing pilot sectors.

For the cantonal housing service

Treat brokers as market sensors: their transaction and demand data, aggregated and anonymised, inform the housing policy chapter 13 calls for. In return, open up clean, up-to-date land registry data, since the valuation tools being deployed will be calibrated on it, and an error in the registers ends up amplified across hundreds of reports.

---

Jérôme Deshaie is the founder of MCVA Consulting SA, an agency specialising in the AI transformation of organisations in Valais, and the author of Bisse Cognitif.