Verify statistics from raw data
Double-check calculations and statistical analyses against the underlying data.
Claude recomputes the reported numbers from the raw data by writing and running code, checks the method fits the question, and flags anything that does not reconcile: a second set of eyes before results ship.
Single model call (rung 2 of 6). Generate, extract, summarize, classify or translate. One request, one answer.
Not an agent: the sequence of steps is fixed in advance, so the whole path can be tested end to end.
- Autonomy: T0 · Suggest
- Surfaces information, ranks, or recommends. Changes nothing. Human role: does all the work; may ignore the system entirely.
- How it is governed
- Output schema validation. A human reads it before it counts.
- Anthropic (Claude Certified Architect – Foundations (CCAR-F))A single, well-prompted LLM call, which is often enough. source
- OpenAI (OpenAI agent guidance (Agents SDK, Codex, OpenAI Academy))One model request with clear instructions and context. source
- Google Cloud (Generative AI Leader)Prompt engineering: zero-, one- and few-shot prompting (exam §3.2). source
- Microsoft (AI-901 (Azure AI Fundamentals) and AB-100 (Agentic AI Business Solutions Architect))System and user prompts for generative AI models (AI-901). source
How Koko classifies AI · the full ladder is taught in AI & Agentic Delivery Foundations.
Try it
Illustrative demoBefore a result goes in a report, someone needs to confirm the numbers reconcile. Edit the inputs to tailor the output — it's baked synthetic data, so nothing leaves your browser until you run it live.
- Recompute these summary stats from the raw data and tell me if the reported figures hold.
- Is the statistical test appropriate for this data, and are the assumptions met?