I asked Perplexity AI who they would recommend for web design for solopreneurs.
My name didn't come up.
Surprising? Not really. I've been self-employed since March 2026 and only recently started implementing GEO optimization properly — that takes time. But the test made something clearer that had been on my mind for a while: anyone who wants to be found online as a tradesperson, landscaper, coach, or consultant in 2026 needs to think on two fronts — not just Google.
ChatGPT, Gemini, Claude, Perplexity — these tools don't respond to search queries with a list of ten results to click through yourself. They deliver a summarised answer, with two to five cited sources. If you're not among those sources, you simply don't exist in that answer.
That's what this article is about:
- Why ChatGPT now answers more questions than Google — and what that means for a tiler, coach, or landscaper
- Why small, carefully maintained websites have the upper hand over bloated agency sites (not a consolation prize — backed by research)
- 5 measures you can implement without a computer science degree
- What an llms.txt file is — and why you want one, even if you've never heard of it
What Has Changed Since 2023 — and Why Does It Affect You?
Google Is No Longer the Only Starting Point
Things used to be straightforward: rank on Google's page one, get found. Google is still dominant — according to a study by Claneo and Appinio among 2,000 German-speaking internet users, 77–80 % of Germans use search engines multiple times a week [2]. That hasn't changed.
What has changed: since 2023 — with the rise of ChatGPT, Perplexity, and later Google's own AI Overviews — a growing share of research no longer starts at google.com but directly at an AI tool. ChatGPT now records over 1.7 billion visits per month (according to estimates by the trade press). Perplexity nearly tripled between March 2024 and May 2025 alone.
For a tiler, this means: when someone asks Claude "Which tiler is there near me in Cologne?" — an AI decides which names get mentioned. Or when someone asks Perplexity "Which coach can help me win more clients as a freelancer?" — the system decides whether your name appears in the answer. Or doesn't.
What Is GEO?
GEO stands for Generative Engine Optimization. The term was coined in 2023 by a research paper from Princeton University [1]. The idea is to structure content so that AI systems select and cite it as a reliable source.
Important: GEO does not replace good SEO. If you're not visible on Google, you have an SEO problem. If you don't appear in AI answers, you have a GEO problem. The two are connected — and both can be improved with the same core measures.
Why GEO Is Worthwhile for Small Businesses and Freelancers Too
Here's the good news — and I'm stating it deliberately, because it often gets lost.
The Princeton University study examined how GEO measures affect visibility in AI responses. Result: GEO can boost visibility in AI answers by up to 40 % [1]. What matters most is content quality — not domain age or backlinks alone.
This is because AI systems weigh different quality signals than traditional search engines. It's not just domain age and backlinks — the questions are: Is the content precise? Does it answer specific questions? Is the site well structured?
A landscaper's website that clearly describes which garden types they design and answers the questions their clients most frequently ask can outperform a bloated agency site here.
5 Measures So Your Website Can Be Found on Google and AI
1. Add Schema Markup
Schema markup is structured data in the source code — it helps search engines and AI systems understand what's on your page. Particularly relevant for tradespeople and coaches: LocalBusiness, Person, FAQPage, and Service.
Websites with correct schema markup are cited more frequently in AI answers.
You can check whether your website has schema markup yourself: right-click on the page → "View page source" → use Ctrl+F to search for application/ld+json. If nothing appears: it's missing.
2. Use Question-Based Headings
Your clients — say, someone looking for a coach — don't type "business coach Frankfurt" into a search bar. They ask: "Which coach can help me win more clients as a freelancer?" Or a landscaping client asks: "What does a new garden layout cost for a terraced house?"
Exactly these questions belong as headings on your website. The goal isn't to convert every heading into a question — but to use question-based headings specifically where your clients would actually ask exactly that.
3. Keep Content Fresh
AI systems favour fresh content. Pages updated within the last 30 days are cited more frequently than outdated pages.
This doesn't mean something new needs to appear every day. One blog article per month already sends the signal: this site is being maintained.
4. Maintain an llms.txt File
An llms.txt file sits in the root directory of the website and tells AI crawlers which content to consider — comparable to robots.txt for traditional search engines, but for AI systems. The standard became established in 2025 and 2026 [3].
Currently, llms.txt is not an official, binding standard, but a useful strategic signal [4]. 68 % of leading AI models already support the standard (Anthropic, 2025).
I set up my llms.txt from the very beginning and have since expanded it significantly — with clear information about my services, my target audience, and my most relevant content. In English, as most AI crawlers prefer this.
If you don't have one yet: it's not a huge effort. But you need to know what belongs in it.
5. Make E-E-A-T Signals Visible
E-E-A-T stands for Experience, Expertise, Authoritativeness, Trustworthiness. This sounds technical, but the core is simple: AI systems ask themselves when evaluating a website — who is behind this, what experience do they bring, and are their statements trustworthy?
In practice:
- Write a genuine about page.
- Describe your experience.
- Name specific references or projects — not as self-promotion, but as context.
What This Means for Tradespeople, Coaches, and Landscapers
The most common mistake I see on solopreneurs' websites: the page describes the business — but doesn't answer any questions.
"We offer professional garden design" is not an answer. "What does a new garden layout cost for a terraced house?" — that's the question someone actually asks. And that question belongs as a heading on the page, followed by an honest, concrete answer.
Write your website texts as if you were answering your clients' most frequently asked questions — not as if you were describing your business.
Perplexity doesn't know me yet. That won't change overnight — AI systems need time to incorporate new GEO signals into their answers. If you want to know what I build into a new website from the start: here's the web design page.
Frequently Asked Questions About Website Visibility on Google and AI
How long does it take for my website to show up on Google?
For new domains, Google typically needs three to six months to build trust — regardless of how well the site is technically built. You can speed things up by submitting a sitemap in Google Search Console and setting up a properly configured robots.txt.
Is GEO optimisation worthwhile for small trades businesses?
Yes — especially for smaller, focused sites. The Princeton study shows that GEO can boost visibility in AI responses by up to 40 % [1]. AI systems weigh content quality more heavily than domain age or backlinks — which benefits well-structured, precise sites.
What does it cost to optimise a website for AI search?
It depends on what's already in place. For a newly built website, GEO fundamentals can be included from the start. For an existing site, it's a matter of effort: llms.txt and schema markup can be added; structural changes take more time.
Do I need a blog to be visible in AI search?
No — but it helps. Fresh content is a positive signal. Even one article per month is enough to signal that the site is actively maintained.
What is the difference between SEO and GEO?
Good SEO optimises for traditional search engines like Google: keywords, backlinks, technical fundamentals. GEO optimises for ChatGPT, Claude, Gemini, or Perplexity to select your content as a reliable source: structured data, clear answers to specific questions, E-E-A-T signals. Good SEO is often a solid GEO foundation too — the measures overlap.
Do I need to implement this myself?
The core concepts are easy to grasp — but the impact often depends on details you only discover once you go deeper: which schema types make sense for which page, how an llms.txt should really be structured, which headings AI systems consider worth citing. Getting this right takes time and experience with exactly these connections.
Get in touch — I'll take a look at what makes sense for you specifically.
Sources:
- Princeton University et al.: “GEO: Generative Engine Optimization” (arXiv:2311.09735, ACM SIGKDD 2024) – collaborate.princeton.edu
- Claneo / Appinio: “State of Search 2024”, 2,000 respondents in DACH – claneo.com
- llms.txt-Generator.de: “llms.txt: Standard for AI crawler control” (2026) – llms-txt-generator.de
- SEO-KÜCHE: “llms.txt explained simply” (January 2026) – seo-kueche.de
