The verdict in three sentences
In Geneva, a villa seller in Cologny or an international buyer looking for a flat in the Old Town uses Google to find agencies, but increasingly asks AI assistants like ChatGPT "which agency really knows the left-bank market?". SEO remains the foundation for searches by neighbourhood and property type, while AEO/GEO determines whether your brokerage is named, and described accurately, in those summary answers. Before splitting a budget, you need a dated diagnostic of what both channels say today.
Two channels, two moments in the decision
Clients of Geneva's prime residential market choose an agency on a few stable criteria: listing quality (properties genuinely under mandate, photos with rights cleared), detailed neighbourhood knowledge, discretion on off-market properties, broker profiles and responsiveness. These criteria are checked differently on Google and in an AI answer.
| Criterion | SEO (Google) | AEO/GEO (AI assistants like ChatGPT) |
|---|---|---|
| Typical queries | "villa for sale Cologny", "apartment valuation Champel", "luxury real estate agency Geneva" | "Which agency can sell a property discreetly in Vandœuvres?", "How do I buy in Geneva if I live abroad?" |
| Moment of decision | Active search for a property or a valuation | Building a shortlist, often before any contact |
| Time to effect | 3 to 8 months on competitive neighbourhood queries | Variable: first changes sometimes within 2 to 4 months, with no certainty |
| Market budget (2026 ballpark) | CHF 2,000 to 6,000 a month with an agency | Audit of CHF 4,000 to 12,000, then CHF 1,000 to 3,000 per quarter of tracking |
| Metrics | Impressions and clicks by neighbourhood, valuation requests, mandates signed by source | Share of tracked questions where the agency is named, accuracy, sources cited |
| Specific risk | Expired listings still indexed, content duplicated from portals | Sold properties or departed brokers still mentioned by AI assistants |
These ranges reflect typical local agency quotes and public rate pages; they vary with the number of neighbourhoods and languages covered.
One point of context: foreign buyers often ask AI assistants about the rules on acquisitions by persons abroad (the federal law known as Lex Koller) or about taxation. A brokerage that publishes clear, careful content on these topics, without standing in for a notary or lawyer, improves its chances of being picked up as a source. This is not legal advice.
What to prioritise for your situation
| Your situation | Where to invest first | Reason | Indicative year-1 split |
|---|---|---|---|
| Recently founded agency, few public mandates | Neighbourhood pages and broker profiles | AI assistants have few reliable sources about you | 70% SEO, 30% AEO/GEO |
| Established brand with a known name in Geneva | AI record and correction of wrong information | You are probably cited; accuracy remains to be checked | 40% SEO, 60% AEO/GEO |
| Mostly off-market business | Expertise content rather than listings | Discretion rules out exposing properties, expertise can still be published | 50%, 50% |
| International clientele (English-speaking, family offices) | English pages and content on buying from abroad | These buyers ask AI assistants a lot of questions | 40% SEO, 60% AEO/GEO |
| Limited budget (under CHF 15,000 for the year) | One audit and 5 key pages | A few solid pages beat many weak ones | 60% SEO, 40% AEO/GEO |
Data rules matter too: valuation forms collect personal data subject to the revised Federal Act on Data Protection (revFADP, in force since 1 September 2023, overseen by the FDPIC), and the GDPR applies as soon as you target EU residents, particularly in neighbouring France.
Measuring without fooling yourself, and what a diagnostic covers
For SEO, Search Console and lead sources in your CRM are enough to track the essentials. For AEO/GEO, the only reliable measure is a dated record: a fixed list of questions, asked in the same words at regular intervals, with answers archived and annotated.
| Indicator | Source | Rhythm | Reading |
|---|---|---|---|
| Clicks on neighbourhood queries (Cologny, Champel, Eaux-Vives, Vandœuvres) | Search Console | Monthly | Slow but cumulative progress |
| Valuation requests by source | CRM | Monthly | The indicator closest to a mandate |
| Agency named across 25 to 40 tracked questions | Dated record | Quarterly | Baseline from the very first record |
| Accuracy (brokers, areas covered, languages spoken) | Annotated record | Quarterly | A repeated error is a priority |
| Sources picked up by AI assistants | Record | Quarterly | Reveals directories and articles to update |
A diagnostic covers, in broad terms: Google visibility on your high-value queries, AI answers to sellers' and buyers' questions, factual errors, the sources feeding those answers, gaps on your site and a recommended split. At Kolonell, it is priced on quote. For reference in real estate, Kolonell built the website of Orven Properties, a real estate agency in Dubai (orvenproperties.com).
Mini case study
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Illustrative example: Maison Salève Immobilier, three brokers, prime residential sales on the left bank, annual digital marketing budget of CHF 36,000.
- Current split: CHF 3,000 a month of SEO, i.e. CHF 36,000, with no measure of what AI assistants say.
- Planned split: CHF 2,200 a month of SEO (CHF 26,400), an AEO/GEO audit at CHF 6,000 and 3 quarterly records at CHF 1,200 (CHF 3,600), i.e. 26,400 + 6,000 + 3,600 = CHF 36,000.
For the same budget, the agency gets a quarterly measure of how often it is named across 30 sellers' and buyers' questions. The record guarantees no recommendation: it shows where to correct and whether corrections have an effect.
FAQ
Can an agency be sure of being recommended by AI assistants?
No. The work targets the sources and content AI assistants draw on, then measures across 25 to 40 tracked questions; any promise of recommendation should be dismissed.
How long before a measurable effect?
For SEO, 3 to 8 months on competitive neighbourhood queries. On the AI side, the first record sets an immediate baseline and changes sometimes appear within 2 to 4 months.
Should we publish off-market properties to be more visible?
No, that would break the discretion promised to sellers. Publish 5 to 10 expertise pieces per neighbourhood instead, showing your market knowledge without exposing any property.
What budget should we plan for a diagnostic?
In the Geneva market, the 2026 ballpark is CHF 4,000 to 12,000 for a one-off audit. At Kolonell, the diagnostic is on quote depending on the number of neighbourhoods, languages and questions.
Do we need an English site as well as French?
For an international clientele, yes: a large share of the questions foreign buyers put to AI assistants are asked in English. The site's languages are set in the quote.
Find out how AI assistants describe your firm. We record what Google and AI assistants like ChatGPT say about your brokerage across your neighbourhoods and property types, then propose a costed budget split. Request a diagnostic quote: reply within 48 hours.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
