
What happens when the AI Agent gets it wrong? Our escalation model
Confidence thresholds, human-in-the-loop, complete logging: Our escalation model ensures no AI agent error goes undetected.
No buzzwords, just hands-on craft: tool comparisons, regional digitalisation and decision support for managing directors who want to get rid of routine.
Articles by Simon Glowik

Confidence thresholds, human-in-the-loop, complete logging: Our escalation model ensures no AI agent error goes undetected.

Process baseline, quick wins, effort estimation: What a two-week Discovery Sprint delivers, and what it deliberately leaves open.

Small Power Apps projects only come with a tiny Dataverse quota. What that costs when it matters and how to prevent it.

n8n can run as an MCP Server for your own AI agents. Setup, transport protocols, and the pitfalls you will hit in production.

When does Power Automate actually store data in the EU? The facts on EU Data Boundary and what it means for your operations.

CISA warns for the first time in 2026 of an actively exploited n8n vulnerability. What self-hosting customers should check and patch now.

Why Copilot reveals SharePoint over-permissions that have accumulated over years, and how to prepare the rollout cleanly with SAM and Restricted Content Discovery.

Wrong order, tight deadlines, full quotas: Why BAFA applications are rejected in 2026 and how to avoid the most common mistakes.

Contract law, liability clauses, escalation model: Who actually bears liability when an AI agent makes a mistake in mid-market settings.