We now know which words Swiss companies use when they search for AI and security, which of them bring money, and which of our pages answer them. It took us one day. The data cost 35 US cents.
Most companies learn this after a year of blogging, when the visits do not come. We wanted to know before we write the next article.
Why we did it
Since 24 October, odcus.com leads with AI, with security as its foundation. A new positioning is a bet that customers think the way we do. We wanted to back that bet with data before building months of content on it.
For this, an SEO tool runs at ODCUS on the same system as the website, Google Search Console and the blog. It is open source and runs on our own infrastructure. We pay per query, there is no monthly subscription. We measured 198 terms on Google Switzerland, plus the competition on the first two pages for twenty target terms.
What we knew after one day
Which words bring money
Swiss companies search very concretely for tools and rules. Monthly searches on Google Switzerland in German, measured in October 2026:
- M365 Copilot: 1,300
- EU AI Act: 1,300
- ISO 42001: 390
- KI-Agenten (AI agents): 210
- KI-Beratung (AI consulting): 140
- Shadow AI: 110
For "Shadow AI", the estimated cost per click in Google Ads is around 27 Swiss francs. Advertisers only bid that high when a click often leads to a purchase. For us, these terms are the front door to customers who are ready to decide.
The cross-check was just as valuable. Terms like "secure AI adoption" or "AI governance consulting" work very well in conversation, but people rarely search for them. We keep that language for conversations and referrals. In search, we use the words customers type.
Where our customers come from
Search Console shows how visitors find us before they know our name. They come in through questions like which data can go into ChatGPT or what ISO 42001 requires. These are security questions about AI tools.
That is exactly where ODCUS stands: bringing in AI, with security as the foundation. The data confirms the positioning we launched on 24 October. Our rule since then: AI leads with the term, security delivers the proof.
Which pages already work
For almost every searched term, we already had the right page. Instead of writing twenty new articles, we sharpened the existing ones: titles in search language, duplicates merged, clean internal links. The new titles for AI consulting and AI agents went live the day after the measurement.
There were two real gaps. We closed the first one the next morning: ChatGPT Team or Enterprise, a topic with clear demand and typical Swiss questions, such as where the data is stored.
Where Google answers by itself
For 17 of 20 target terms, Google shows an AI overview above the results. Whoever gets cited there wins the reader before they scroll. That is why our pages answer the searched question right at the top, clearly and in a form that can be quoted.
What we get from it
At ODCUS, the website, Search Console, keyword data and the blog run on one system. When a post is planned, the search term and the page that owns it are already set. Every week, the system checks where we rank for thirty terms and records what moves.
Nobody opens a dashboard for it, and between measurement and change, nobody copies a spreadsheet from one program into the next. What companies used to buy as an audit with a PDF and a list of actions now runs at ODCUS every week. What this looks like in the rest of our work is on AI at ODCUS.
What you get from it
Take an accounting firm that blogs about "digital bookkeeping" for a year. Meanwhile, its customers type "Bexio or Abacus" and "VAT return deadline". The content is good, it just sits under the wrong word. One measurement before the first article would have saved that year.
Every company that measures before it writes gains three things:
- Budget that reaches buyers. Every article targets a term that customers with buying intent search for.
- Existing pages work harder. A sharpened title often brings more than a new post.
- Knowing instead of guessing. Every week it is clear what works, without anyone pulling numbers together.
Measuring costs cents today. The difference is a system where the numbers arrive where the content is made, and someone acts the same day. If your website, search data and content still live in three places, let's look at it in a 30-min first call.
Do you know what your customers search for before they know your name?
Frequently asked questions
How much does keyword research cost?
The data itself is cheap today. Search volumes are paid per query; our measurement of 198 terms for Switzerland cost 35 US cents. The expensive part is not the measurement but the interpretation: which term belongs to which page and what has to change on the website.
Do I need an SEO subscription like Ahrefs or Semrush?
Not necessarily. There are open-source SEO tools that pull search volumes, rankings and backlinks per query instead of through a monthly subscription. For an SME that measures a few times a year and tracks a few dozen terms weekly, that is usually enough.
Why is my page not found even though the content is good?
In our experience, it is rarely the text. Often several of your own pages compete for the same term and split the attention, or the title does not use the word customers actually type. The third reason is authority: other websites do not link to the page often enough yet.
