Blog
Explore articles, guides, and resources about data engineering, AI, and analytics.
Explore articles, guides, and resources about data engineering, AI, and analytics.
Keboola connects to every entity's ERP, standardizes the chart of accounts, eliminates intercompany transactions, and feeds Power BI governed, consolidated numbers. No rip-and-replace, first entity live in 8 weeks.
Keboola standardizes financial data from every portco's ERP into one reporting structure, then feeds board-ready, audit-ready reporting. No rip-and-replace, first portco live in 8 weeks.
Anthropic's research shows Claude behaves differently by language. Jordan analyzed around 3,000 internal Kai conversations to find out if Kai does too. The answer: yes, and the differences are bigger than expected.
The most expensive mistake in finance technology is buying the right platform for the wrong problem. This guide tells you which layer you are actually missing.
Finance teams are choosing static data over live AI, on purpose. Here is why that is rational, and what it tells you about where AI in finance actually is.
When the job description says analysis and the actual job is seventeen spreadsheets, the best people leave. Here is why retention interventions keep missing the point.
Getting the data right is the straightforward part. Deciding what the organisation becomes once it is right is harder. That decision belongs to the CFO.
Most finance leaders have now tried an AI assistant on real data. Most of them had a similar experience. Here is what is actually happening, and what trustworthy finance AI looks like when it works.
The deadline that catches most incoming PE CFOs off guard, and the 8-week framework to make sure it does not catch you.
A CFO asked Claude for the year-end cash position. The answer was 30% off. The problem was not the model. It was what the model was reasoning over. Here is what ungoverned financial data does to AI, and what a governed layer changes.
FP&A teams spend 70% of their time collecting data, not analysing it. The rolling forecast is stale not because the team is slow, but because the data pipeline feeding it was never built for speed. Here is where it breaks down, and what changes when it is fixed.
Five ways manual account mapping fails after an acquisition, and what finance teams are doing instead.