Blog

I only write about what you can put to use directly: proven workflows, ready-made prompt templates, numbers from real projects and conclusions I have drawn myself. Most articles are published in German first.

19 Jul 2026 Why visibility in one AI engine says little about the next: one month of retrieval-to-citation data from eight B2B projects
I spent a month measuring how often ChatGPT, Perplexity and Google AI Overview actually cite a source they retrieved. Across eight B2B projects the pattern holds: three engines, three near-separate source pools. AI visibility is three targets, not one.
17 Jul 2026 Loop engineering: what the new way of working with AI agents is worth — and when a good prompt is enough
Loop engineering has AI agents work in cycles until a stop criterion takes hold. Which loop types exist, why loops fail in practice, and which seven questions to answer before your first own loop.
15 Jul 2026 A skill knows nothing: brand knowledge as an OKF bundle instead of a pile of files in the chat
Skills execute, they know nothing. How to store brand knowledge in the Open Knowledge Format: a folder of Markdown files any model reads without conversion. With my complete GEO bundle to download under CC BY 4.0.
5 Jul 2026 The prompt as a Lego figure: why AI visibility runs on entities, not keywords
AI systems break every prompt into several sub-queries and search for answers in parallel. Whoever optimizes only for the starting term is absent from the majority of these runs. Why entity mapping replaces the foundation that keyword thinking once was.
3 Jul 2026 llms.txt: what the file does, what Google says about it — and when you should create one
The complete guide to llms.txt: origin, structure, Google's official position, what the data shows about its use and who the file really pays off for — with my own analysis of 2,041 source domains, an example and practical steps.
3 Jul 2026 Not every rule belongs in the prompt: steering AI instructions and model memory
Six zones decide where an AI instruction belongs — from the product settings to the current prompt. How to instruct ChatGPT and Claude cleanly and keep their memory of you current.
29 Jun 2026 Google ends the June spam update in two days — and the forums report more spam, not less
The June 2026 spam update was over in two days. Why the loud complaints about traffic drops and rising spam are not yet a data picture, and what you should really check instead of forum panic.
29 Jun 2026 How ChatGPT chooses its sources — and the GEO levers that follow
A look at the traffic beneath the answer shows how ChatGPT gathers, selects and cites sources. What of it is actually evidenced, and where to act for your AI visibility.
27 Jun 2026 When the AI answers too shallowly: eight techniques to make it think deeper
Strong prompts are not magic formulas. They give the model a mode of thinking. Eight techniques to open up a topic and secure a decision — with ready-made prompt templates.
25 Jun 2026 Retrieved but not cited: the most valuable gap in your AI visibility
When an AI model retrieves your page but does not cite it, that is your biggest opportunity. How to measure the gap, prioritize it, and track it over time.
23 Jun 2026 The system prompt, taken from Claude Fable 5 — and which prompting techniques you can adopt from Anthropic
The complete 17,000-word system prompt was pulled from Claude Fable 5. What a system prompt even does, how far you can steer a model's answers — and eight techniques Anthropic builds into its models.
17 Jun 2026 GLM-5.2: the strongest open model is text-based and token-hungry
Z.ai released GLM-5.2 as open weights under an MIT license. In the independent Intelligence Index it leads the open field. Two caveats decide the practical value: it takes only text and consumes a strikingly high number of tokens.
12 Jun 2026 When your AI cuts corners: five techniques to get more out of every answer
Even a well-considered AI answer is often raw. Five techniques bring answers up to working quality, uncover hidden weaknesses in your ideas, and can be linked into chains.
1 Jun 2026 Where the tokens go: steering the context of AI models
The context window is finite, and the fuller it gets, the worse the model works. How to make the context visible, know its working zone and place a task into it without wasting quality and tokens.
10 May 2026 How Claude Code is built: memory, MCP, skills, subagents, hooks
Claude Code looks like a command line from the nineties. Beneath it lie six layers that make the tool a construction kit. A look under the hood, with examples from my work.
12 Aug 2025 Google Search Central APAC 2025: what really counts in SEO practice
A sober look back at Google Search Central Live APAC 2025 — the points that make a measurable difference in day-to-day SEO.
30 Jun 2025 Crawl budget: how HTTP status codes affect your SEO ranking
Why HTTP status codes govern crawl budget and how clean status-code management noticeably raises crawling efficiency.
18 May 2025 Why keyword research can distort your SEO strategy
Classic keyword research often optimizes for volume instead of value and thereby distorts the whole strategy. How to avoid it.
9 Apr 2025 Pricing first: why Germans watch prices
In the German market, price is often the entry point to the buying decision. What that means for dynamic pricing strategies, platform presence and value communication.
21 Mar 2025 The relevance of LLM visibility for online shops
Buying decisions are shifting into AI answers. Why LLM visibility becomes a channel of its own for online shops and how to occupy it.