Knowledge
In-depth articles
Guides, tutorials and background knowledge on generative AI in patent practice.
Office + AI Integration
VBA macros and Python bridge for Word & Excel with LLM connectivity (Ollama, ChatGPT)
15 min readRAG Workshop Guide
Search your own documents with AI — local RAG approach with ChromaDB & Ollama
20 min readClaude Code Guide
Comprehensive reference for Claude Code — terminal-based AI development
30 min readLLM Overview
Comparison of current language models — capabilities, pricing, use cases
10 min readInteractive Visualizations
Token Sampling, Dense vs. Mixture of Experts — interactive demos to understand LLMs
5 min readMixture of Experts (MoE)
MoE architecture explained — how it works, comparison with dense models, practical relevance
10 min readVector Search & Embeddings
Semantic patent search, hybrid search, and RAG: what AI finds that keyword search misses, and the limits to know before deploying it
10 min readPrompt Engineering for Local LLMs
Patterns and anti-patterns for gemma3, qwen, llama, what is different with local models
20 min readDuty of Evidence for AI Answers
Every AI statement backed by a verbatim quote with reference. The rule that prevents hallucinations in patent work.
7 min readParagraph-by-Paragraph Review
Systematically through the entire document, not just the model's intuition. A pattern for long patent specifications, office actions and prior art.
7 min readAI Agents & Isolated Context
What defines an AI agent and why isolated context is the central feature
7 min readMCP · Connecting AI Models to Patent Data
How AI assistants reach into patent registers via the Model Context Protocol, walked through with the open patent-client-agents example
12 min readThe Anthropic IP Plugin — Blueprint for IP Work
What the open-source plugin claude-for-legal/ip-legal solves cleanly, what it deliberately does not, and which eight design patterns transfer to an EPC/EU practice
14 min readYour Own IP Suite as a Claude-Code Plugin
Wayfinder: how to assemble your own European and German patent suite (search, drafting, prosecution, opposition, litigation) from skills, MCP servers and output discipline, across cloud, EU-cloud and local AI paths
13 min readContext Engineering in Patent Work
Evolution of prompt chaining, for drafting, office-action response and FTO that juggle many documents
12 min readPrompt Engineering Patterns for Patent Work
Eight recurring building blocks: Role-Task-Format-Context, chaining, chain of thought, few-shot, self-reflection, devil's advocate, fact-checking, structured output
10 min readMinto Pyramid and Autoregressivity
Why good technical writing only emerges in revision, why language models structurally cannot revise, and why iteration is the only compensation for it
14 min readMulti-Model Workflows
Why two AIs catch more than one. Drafter and reviewer as a recurring pattern in patent practice
8 min read4 Levels of AI Practitioner
A competency framework from occasional user to tool maker, with a 30-second self-assessment and a matching next step
10 min readUsing the Patent Description Generator
Step-by-step guide to the Description Drafter — inputs, variants, prompt tabs, workflow
8 min readVideos
Videos about AI — talks, demos and background knowledge
Adopting local AI in a patent firm
Local AI for smaller patent firms — hardware paths, software stack, where the cloud advantage shrinks and why local inference is enough for drafting.
8 min readAssistance or integration
Two operating modes for the same toolbox: AI inside the existing work step, or a workflow rebuilt around the very same tools. What that shifts within the term and what stays the same.
9 min readGenAI Tool Collection
Link collection — software tools for generative AI in patent work
Live
Workshops and events
My GenAI Workshops at Forum Institut
Current e-seminars on generative AI in patent practice, live, with hands-on examples and time for questions.
View seminarsPATINFO 2026, talk on 12 June 2026
Session „Patentinformation & Recherche", 11:15: „KI-Sprachmodelle als Recherche-Assistenten: Zwischen Halluzination und Effizienzgewinn" (talk in German).
View eventReading
Book recommendation
Patent Portfolio Management
A Practical Guide in the Age of Artificial Intelligence, Second Edition
Simone Frattasi, Jean-Claude Alexandre Ho · Edward Elgar Publishing
View book