Guides on AI analytics and business intelligence
Practical, sourced and dated. We say where a fact comes from, which sources are vendors, and what a test or approach cannot prove.
Guides
Plain explanations and how-tos.
- What is agentic business intelligence?Agentic BI lets AI plan and run analysis toward a goal instead of answering one query at a time. What it is, how it differs from dashboards, and its limits.7 min read
- How to chat with your SQL database — and trust the answersAsk your SQL database questions in plain English. How it works, how to give a tool safe read-only access, and how to check every answer before you rely on it.8 min read
- Natural language to SQL: how it works, where it fails, and how to use it safelyHow natural-language-to-SQL works, why schema linking is the hard part, how accuracy is measured, and how to use it safely on a real database.9 min read
- Conversational BI vs traditional BI: what's the difference?Dashboards and conversational analytics answer different questions. When each fits, where they overlap, and why most teams end up using both.6 min read
- What Power BI Copilot requires before you can use itPower BI Copilot needs a paid Fabric (F2+) or Premium (P1+) capacity and an admin to enable it. The requirements, limits, and when a simpler tool fits.6 min read
- How to automate business reporting without losing trustStop rebuilding the weekly report: choose what to automate, set up live dashboards, scheduled snapshots and alerts, and keep every number checkable.7 min read
Original research
What we tested, found and fixed — with the method.
- How we test an AI data analyst: 115 questions, independently computedWe check every release against 115 questions with independently computed answers. The method, the failures it caught, and what the test can't prove.8 min read
- Why your spreadsheet isn't one table: messy-data traps that break totalsStacked sections, pasted lists and second headers quietly break totals. A worked example of how messy files fool analysis tools, and how to check yours.6 min read
Free resources
Checklists and kits you can use with any tool.
- The 10-question accuracy test for any AI analytics toolA free, tool-agnostic test: ten question types, how to compute each answer independently, and a scoring sheet. Run it on any AI analytics tool, ours included.6 min read
- Is your data ready for an AI analyst? A 12-point checklistTwelve quick checks that prevent wrong answers from AI analytics tools: headers, stacked sections, blanks, dates, duplicates, units, and more. Free checklist.5 min read
- 24 weekly questions for operations, retail and finance teamsEight weekly questions each for logistics operations, retail and e-commerce, and finance: the data needed, what a good answer shows, and the trap to avoid.7 min read