System StatusOnline
NAV
Ruminations on code, intelligence, and the architectural voids of the modern web.
Chatbot, AI agent, or human escalation — this decision costs companies a fortune when made wrong. A decision matrix, three case studies, and five deployment mistakes that destroy CSAT instead of improving it.
An employee spends 3 hours daily re-typing data from invoices, contracts and forms. At 17 EUR/hour that's 4,600 EUR per year — for one process. I show how to build an OCR+AI pipeline with 90%+ accuracy and who's accountable when the model makes mistakes.
Most companies deploy AI because they feel they have to, then measure success by whether the system runs at all. I show you how to calculate ROI before deployment — four numbers, one calculator, three case studies with real figures.
The agent was given a task, processed data for 40 minutes and finished nothing. Diagnosis: this is a job for 4 agents, not one. I show you how to recognise when you need a multi-agent system, how to design it, and what mistakes to avoid when deploying.