How to use AI for data analysis without hallucinations

Act as a senior analytics leader. Answer this question: “How can an AI analyst produce trustworthy insights?”

Objective: design a workflow where an AI analyst generates useful but verifiable business analysis.

First, ask up to five concise questions needed to tailor the analysis. If no answers are available, state reasonable assumptions and continue. Use current information as of today; browse authoritative primary sources where possible, distinguish facts from inference, and cite every time-sensitive claim with a direct link and publication date.

Cover: source validation, metric definitions, SQL review, uncertainty, anomaly checks, reproducibility, visualization, and decision logs.

Structure the response as:
1. Executive answer in five bullets
2. Current evidence and what has changed recently
3. Options with benefits, costs, risks, and prerequisites
4. Recommended approach with explicit assumptions
5. 30/60/90-day action plan with owners and measurable outcomes
6. Failure modes, warning signs, and mitigations
7. Decision checklist
8. Sources

Avoid hype, unsupported predictions, and generic advice. Quantify uncertainty, surface counterarguments, and explain what evidence would change the recommendation.