Blog
Insights on custom AI agents and enterprise automation.
The Knowledge Half-Life: Why Your Wiki Is Already Out of Date and a Company Brain Never Is
Static wikis, SharePoint, and SOPs decay from the day they are written - they have a knowledge half-life, because nobody updates them and they never observe the work. Why a company knowledge base built on AI stays current: a Company Brain is a living memory fed by daily work and feedback that survives turnover instead of walking out the door.
Henri Jung · 33 min read 
Hiring Freeze, Not Layoffs: How to Grow Output While Headcount Stays Flat
A deliberate hiring-avoidance strategy for the Mittelstand, run department by department with AI employees: how to grow output while headcount stays flat, without mass layoffs. Covers the output-per-person maths, what to freeze and what to keep hiring, the Company Brain that makes the gains durable, the Klarna over-cut cautionary tale, DSGVO and the EU AI Act, and a 90-day playbook.
Henri Jung · 34 min read 
The AI Employee on the IT Service Desk: Closing Internal Tickets Before a Human Sees Them
How an AI employee closes routine internal IT tickets end to end - password resets, access requests, provisioning, and first-line triage - grounded in the Company Brain and connected to your ITSM, Teams, and identity provider. Covers the ticket-cost economics, why it works in 2026, security and DSGVO/EU AI Act guardrails, a 90-day playbook, and how it differs from a bundled ITSM chatbot.
Henri Jung · 34 min read 
Can You Trust Your AI Employee? Observability and Evaluation for Agents in Production
How companies watch, measure, and prove an AI agent's behaviour once it is live: the four layers of observability (traces, evaluation suites, drift alerts, audit trails), offline and online evals, LLM-as-a-judge, EU AI Act Article 12 record-keeping, a 90-day instrumentation playbook, and how the same instrumentation that earns trust also powers the Company Brain feedback loop.
Henri Jung · 36 min read 
AI Employees in Customer Success: Turning Renewals and Expansion Into a Managed Motion
A practical guide for CS and revenue leaders on deploying an AI employee that owns the routine customer success work end to end - account health scoring, renewal prep, churn-risk signals, adoption monitoring, QBR drafting, expansion spotting, onboarding follow-through, and CS-ops reporting - all connected to your CRM, email, Teams, product data, and ticketing. Covers the retention economics, the Company Brain that survives CSM turnover, DSGVO and the EU AI Act, a 90-day playbook, and how it differs from a CS dashboard and a support bot.
Henri Jung · 34 min read 
Agent Washing: How to Tell a Real AI Employee From a Rebranded Chatbot Before You Buy
A buyer's guide to agent washing in 2026: five concrete tests, a demo script, and twelve vendor questions to tell a real AI employee - persistent Company Brain memory, write access to your real systems, learning from feedback, end-to-end ownership - from a rebranded chatbot before you sign.
Henri Jung · 33 min read 
The Feedback Loop: How Your AI Employees Get Better at Your Company Every Week
How an AI feedback loop turns daily corrections, examples, and preferences into a Company Brain that makes your AI employees better every week - and why a static chatbot never improves. Covers the mechanics of the loop, why it compounds into a moat competitors cannot copy, how to operationalise it in a mid-sized company, and how to measure it.
Henri Jung · 34 min read 
NIS2 Meets AI Agents: Securing Agent Deployments Under Germany's Now-Binding Cybersecurity Law
AI agents are a new attack surface and a newly BSI-regulated asset at once. How the Mittelstand deploys agents under Germany's binding NIS2 law: ISMS scope, Section 30 risk management, the 24/72/30 incident clock, supply chain security, and personal management liability - with a 90-day NIS2-ready playbook.
Henri Jung · 34 min read 
Small Language Models on EU Soil: The On-Prem Agentic Architecture the Mittelstand Has Been Waiting For
How the SLM-router pattern lets small models on EU soil handle 80-90% of agent steps, escalating only the hard cases to a frontier model. Covers the 10-30x cost saving, the CLOUD Act and DSGVO sovereignty dividend, model choices (Qwen3, Phi-4, Llama 3.2, Nemotron Nano 2), three deployment topologies, and a 90-day playbook.
Henri Jung · 32 min read