AI for Analysts

Use AI as an analyst. Not as a user. LLMs, agents, and RAGs are changing analytical practice. Learn how to integrate them with evidentiary rigor, audit the output, and design workflows that introduce no risk—instead of promising miracles. VALUE PROPS • Prompting tradecraft: structured prompting, chain-of-thought, role separation, calibrated few-shot. • AI-augmented OSINT: RAG, web search, fact-checking, hallucination management, verifiable citation. • Risk and governance: automation bias, prompt injection, data leakage, LLM trust calibration. • Agentic workflows: human-in-the-loop patterns, audit trails, integration with Claude Agent SDK and MCPs. «WHAT YOU WILL BE ABLE TO DO» • Design analytical prompts with auditable structured outputs • Build a RAG workflow on proprietary sources with verifiable citation • Identify and mitigate prompt injection in real pipelines • Apply SATs (Module 4 of Course 1) to the output of any LLM before integrating it into decision-making • Distinguish legitimate agentic from automation theater