Blog articles
Notes on automation systems, n8n workflows, AI agents, integrations, content pipelines, and the logic behind making messy processes work reliably.
Half of audit work is cleaning data nobody cleaned
May 25, 2026
Alteryx, Power Query, Python — the tool matters less than the fact that someone's data is always messier than expected. Cleaning it is the work.
The slippery slope nobody warned me about
May 24, 2026
Power Query for one report. Then SharePoint. Then Power BI. Then APIs. The accounting automation journey rarely starts where people think it ends.
The breaking point was entering the same invoice twice
May 23, 2026
Manual invoice entry isn't just slow — it accumulates errors that cost more time to find than the entry ever saved. The mental fatigue point is underrated.
The best AI for accounting isn't one tool
May 22, 2026
After testing AI tools for invoice processing, categorization, and anomaly detection, the conclusion is the same: it's the combination, not the model.
300 invoices a month. That's not an AI problem.
May 21, 2026
At 5-10 minutes per invoice, 300 invoices a month is 25-50 hours of manual entry. The part most people skip is building the GL mapping table.
Your bookkeeping software is an automation decision
May 20, 2026
Switching from e-Boekhouden to Moneybird isn't just a UX preference. It's a choice about what you can automate later, and what stays manual forever.
The accountant you need once a quarter
May 19, 2026
A bookkeeper's advice for Dutch starters: do your own admin, use the bank connection, and bring in a professional for a quarterly review.
The Cheapest Model Does the Most Important Job
May 18, 2026
A student built an n8n invoice controller where a cheap model as the first filter turned out to be the most valuable node in the workflow.
The Week Before Building Beats the Month of Building
May 17, 2026
Most automation fails not from bad code, but because the builder started too early — the discovery work is the hard part.
The assembly line nobody planned
May 16, 2026
An accountant who automated his way through PwC, Python, and ChatGPT — and the question it left: why did I spend years studying to end up doing this?