The Lab
Applied AI
13 articles on AI in UX testing, GEO, prompting, voice AI and more.
What do you review when you cannot review everything?
56,198 public pages from 84 German statutory health insurance sites, machine-triaged. Which review patterns recur across insurers, and how editorial teams turn them into an order of work.
Read articleFOMO Filter for AI News in Detail
I've been running a weekly AI briefing agent for months. Four hours to build, a few cents a month. Here's what actually works and what I had to rewrite.
Read articleVoice AI with Gemini Live: Prototype Set Up in 3 Hours
A bad screening call was the trigger. This prototype shows what Google's Gemini Live API can do today — for cents per call.
Read articleControlling AI Understanding with Schema and llms.txt
Schema markup helps Google but LLMs can't read it. What actually determines whether AI cites your content correctly?
Read articleHanzi Movie Method: Learn Chinese Characters with AI
How the Hanzi Movie Method turns Chinese characters into film scenes, and how a free browser app writes the mnemonic scripts for you. No sign-up, English and German.
Read articleAI Progress: You Don't Need to Keep Up With Everything
The flood of AI releases creates pressure to stay current at all costs. Three filters help you ignore the noise and focus on what actually matters for your work.
Read articleAI Overviews Are Stealing Your Clicks: A GEO Playbook
When AI Overviews answer queries before users click, rankings alone do not protect traffic. This playbook shows how GEO can recover visibility and demand.
Read articleClaude Humanizer Skill: Make German AI Text Sound Human
Free Claude Code and Codex skill that makes German AI text sound more human. No Python required, with 72 writing patterns and factual safeguards.
Read articleVerbalized Sampling: More Diversity Without Fine-Tuning
Verbalized Sampling (VS) is a prompt-based trick to fight Mode Collapse in aligned LLMs—getting you more creative answers without any training.
Read articleAI Chatbots in UX: Why Reliability Beats Showmanship
Users trust chatbots that are clear, honest, and useful. This article shows what reliability in UX tests looks like and how to design for it.
Read articleTaxonomy in AI Prompting
How a fixed category structure (taxonomy) makes AI answers comparable, machine-readable, and reproducible, with a JSON example and clear limits.
Read articleWhat AI Reveals About Product Management
AI automates product artifacts but highlights what still matters: judgment, synthesis, and deep context. Learn where human product leadership adds value.
Read articleWhen LLMs Become the Browser: What Changes for the Web
If navigation shifts from links to answers, reach, discoverability, and brand presence all change. What that means for the web and for visibility.
Read article