Blog
    SEO, GEO & LLM
    September 25, 2026· 7 min readLire en français

    AI Visibility Benchmark 2026: Swiss SMEs Measured

    Birdy measured the AI visibility of 112 Swiss SMEs. 73% are absent from the searches ChatGPT consults. Here is what our data actually shows.

    AI Visibility Benchmark 2026: Swiss SMEs Measured

    Between August and September 2026, Birdy analysed 112 businesses in French-speaking Switzerland: landscapers, estate agencies, trades, consultancies, retailers and B2B providers. For each of them, the same question. When a potential customer searches for your service in your region, do AI engines find you?

    The answers are measured, not estimated. They show a clear gap between what Swiss SMEs have already built and what they could get out of it.

    What we measured, and how

    Every analysis follows the same method.

    1. We list the searches your customers actually type: your trade, your town, your specific services. For example "landscaper Montreux", "garden maintenance Riviera", "landscape masonry terrace Montreux".
    2. We query the indexes AI engines rely on: Google, Bing, Brave, plus brand searches.
    3. Search by search, we record your position and the businesses that take your place when you are absent.
    4. We keep two numbers apart: your measured presence (the searches where you show up) and your readiness (the signals that make your business readable and citable).

    Nothing is extrapolated. A search we could not test stays untested, a missing signal is reported as missing.

    The result in one sentence

    Swiss SMEs have a solid technical base and weak AI visibility: 12% of the searches tested on average, and 5 of the last 19 analyses at zero.

    The technical base is rarely the problem. Across 93 detailed analyses, the average technical score is 92.5 / 100 and 92 of 93 sites run over HTTPS. Content tops out at 45.8 / 100, and machine readability for AI engines at 30.9 / 100.

    Distribution of readiness scores across 93 analyses.

    SituationBusinessesShare
    Below 30 / 10044%
    Between 30 and 594144%
    60 and above4852%

    Average 60.6 / 100, median 60, extremes 17 and 92. Almost nobody is badly prepared, and almost nobody is ready.

    Engine by engine

    Assistants do not answer out of thin air, they rely on indexes we can query.

    Source consultedAverage scoreBusinesses absent
    Google index (Gemini)42.9 / 10014 of 85
    Bing index (ChatGPT)14.5 / 10062 of 85
    Brave index (Claude)23.0 / 10024 of 59
    Perplexity index51.9 / 1006 of 26

    How to read this: 73% of the sites tested appear in none of the searches ChatGPT consults. That is the most common blind spot, because almost nobody works with Bing. A business can lead on Google and still sit outside the reach of part of the assistants.

    These measurements cover indexes, not the answers written by the assistants. Appearing is a necessary condition, not a guarantee of being cited.

    The four blind spots

    1. A missing or neglected Google Business Profile. 20 of the 67 sites tested have no Google Business Profile at all. Only 35 of 93 embed a map. The average "Google reviews" score is 57.1 / 100 and 7 businesses have no online reviews. It is the simplest signal for an engine and for a customer to check.

    2. Content too thin to be reused. 19 of 93 sites offer real service pages, 36 of 93 publish a blog. AI engines cite what explains, quantifies and compares. A 300-word homepage gives them nothing to reuse.

    3. Machine readability is almost absent. 13 of 93 sites declare their information with structured data (schema.org). 23 of 93 show a FAQ. 0 of 93 mark it up so an AI can attribute the answers to your business. This is the most profitable gap to close: the work is one-off and it serves Google and the assistants at the same time.

    4. Hard proof is still rare. 49 of 93 sites publish concrete facts (indicative prices, lead times, volumes, results). 44 do not. 68 of 93 display an address. An engine that finds no price, no lead time and no service area has to guess, and it rarely cites a business it has to interpret.

    Reading by sector

    SectorAnalysesAverage score
    Landscaping and outdoor design1451 / 100
    Estate agencies564 / 100
    B2B marketing and strategy478 / 100

    Landscapers show the most common pattern: a solid business, a clear service area, dense local competition, and near-zero presence in the searches ChatGPT consults. Estate agencies do better, carried by listings refreshed continuously. Marketing firms apply to themselves what they sell.

    These comparisons stay directional: each sector counts few analyses, they point to a trend, not to a ranking.

    How to read your own score

    • Below 30: AI engines rarely recommend you. Your prospects get an answer that talks about your competitors.
    • Between 30 and 64: you exist on some searches, not on the profitable ones. Several opportunities remain open.
    • 65 and above: your presence is already good, the challenge becomes consistency and defending that position.

    Two complementary readings. High readiness with weak presence means you are readable but not picked up: the external mentions AI engines consult are missing. Decent presence with weak readiness means your position rests on very little, and a structured competitor can overtake you quickly.

    The first 30 days

    Week 1: stop being invisible locally. Create or fix the Google Business Profile, make name, address and phone consistent everywhere they appear, ask your ten most satisfied clients for a review.

    Week 2: make one service readable. One page per service, with the problem solved, the service area, an indicative price or range, a lead time, one real example. That is the format AI engines reuse.

    Week 3: become citable. Add a FAQ of eight to ten questions taken from the field, then mark your information up with structured data.

    Week 4: measure again. Compare search by search, keep what works, remove what produces nothing.

    You can run these steps in house or hand them to us: our Agency universe explains how we build visibility and acquisition systems, and our SEO support covers the search side.

    Run the same measurement on your business

    The Birdy simulator computes the same indicators for your business in under a minute: your score, the searches where you appear, the ones where a competitor takes your place, and the priority actions. Free, no commitment, no phone call.

    Frequently asked questions

    What is AI visibility?

    Your business's ability to show up in the answers of assistants and conversational engines (ChatGPT, Gemini, Claude, Perplexity, Copilot) when a prospect searches for your service in your region.

    How does Birdy measure it?

    By querying the indexes those engines consult, using the searches your customers really type, then recording your position and the businesses that replace you. Nothing is estimated when it could not be read.

    Should we drop SEO then?

    No. Search remains the main entry point in Switzerland, and the indexes the assistants use are largely the same. The difference is that a page now also has to be citable, which means structured and factual.

    How long before an effect?

    Technical fixes and the Google profile show effects within weeks. Presence in generated answers follows the engines' re-indexing rhythm: expect one to three months for a visible shift, and regular tracking so you are not steering blind.

    Can small providers get cited?

    Yes. Assistants compare options, they do not pick the biggest budgets. A trades business with clear pages, reviews and concrete numbers can be recommended ahead of a competitor that spends far more on ads.

    Who is Birdy Consulting?

    Birdy Consulting is a marketing agency based in Switzerland that helps SMEs with visibility, acquisition and conversion. It measures the AI visibility of its clients and builds the systems that make them findable and recommended.

    Method and limits

    Sample. 112 businesses in French-speaking Switzerland analysed by Birdy in August and September 2026, including 93 full analyses. These are businesses that asked for an analysis, not a random panel of the Swiss economy.

    Measurement. Presence in the indexes assistants consult, on the searches tested, plus the observable signals on the site. We do not capture the answers written by the assistants, which vary with timing, phrasing and region.

    Interpretation. These are observed averages, not performance guarantees. An engine may cite a business absent from our measurements, and the other way round.

    Go further

    Free diagnosis

    Check if your business is recommended by AI

    In under one minute, get your Birdy Visibility Score and the priority actions to improve your visibility on Google and AI engines.

    Get my score

    Monthly marketing insights

    One email per month. Concrete tactics for Swiss SMEs. No spam, unsubscribe in one click.

    Need marketing support?

    Let's discuss your challenges and find the best strategy for your SME.

    Let's Talk

    Related articles