Skip to content

Learning path 03

AI applied to intelligence

Find out what to delegate, how to verify results and how to build reproducible instructions.

≈ 2 h 35 min · Beginner to intermediate

The question it answers

How do you use AI without mistaking fluency for evidence?

What you will take away

A delegation and review protocol for producing reproducible analyses.

Start with 3.1

The Prompt Factory generates instructions for another AI; it does not analyse documents or generate final products.

Five lessons

Recommended order. You can start with any of them.
  1. 3.1What generative AI can and cannot doWhich part of the analyst's work can we delegate to AI, and which part should we not delegate without controls?≈ 25 min · Beginner · You take away a "task / autonomy / evidence / validation" sheet that you can add to your monitoring plan
  2. 3.2Researching with AI in a verifiable wayHow do you turn AI-assisted research into a file that someone else can audit?≈ 35 min · Intermediate · You take away brief evidence file and a list of prioritised gaps
  3. 3.3Classifying and structuring informationHow do you turn heterogeneous documents into comparable records without hiding classification errors?≈ 30 min · Intermediate · You take away data dictionary and disagreement log
  4. 3.4Analysing relationships and producing intelligenceHow do you move from a collection of records to strategic hypotheses without turning correlations into facts?≈ 35 min · Intermediate · You take away hypothesis matrix and draft briefing with uncertainties
  5. 3.5Agents, tools, persistent knowledge and interoperabilityWhat is the difference between asking an assistant, running a task with tools and sustaining a corporate intelligence system?≈ 30 min · Intermediate · You take away architecture diagram and governance checklist

Lab practice · L05

Evidence auditor

Output: Claims classified as fact, inference, error or no evidence. No sign-up required.

Learning paths