FreeAI BasicsPromptingRAGAgents

From Asking to Using AI: AI Application Basics

Move from casual prompting to intentional AI task design

A task-based beginner course that helps you turn a real need into an executable, checkable, and improvable AI usage plan.

适合人群

Beginners who use AI often but get unstable outputs
Professionals who want to apply AI to real work tasks
Learners who want to understand how Prompting, RAG, Agents, and Evals fit into practice

Your requests are too vague, so AI outputs often drift
You are unsure when to add context, sources, formats, tools, or evaluations
You have heard many AI concepts but do not know how to apply them to your own task

Rewrite vague requests into clear AI task instructions
Design context, output formats, source boundaries, and guardrails for a real task
Complete a submit-ready AI usage plan

Carry one real need through the whole course, from task instruction to final AI usage plan
Place Prompting, RAG, tool calling, Agents, and Evals inside one executable workflow
Produce part of the plan in every lesson, then assemble a submit-ready and reusable final artifact
Train when to add sources, formats, tools, and guardrails instead of stacking AI buzzwords

9 节课

Identify a real work or learning task that is suitable for AI practice, and describe its context, desired result, and current pain point.

A short real AI task description. You may type it directly or use voice input and submit the transcript.

Rewrite a vague request into an AI task instruction that includes background, goal, materials, constraints, format, and checking requirements.

The original vague request, the rewritten prompt, and the reason for each key addition.

Decide which context is useful for the task and turn the AI output into a copyable, checkable, reusable format.

A context checklist, a structured output template, and an explanation of necessary versus distracting information.

Judge whether an AI task must be answered from files or source materials, and define source scope, retrieval focus, and no-fabrication boundaries.

A source-dependency decision table covering whether sources are needed, the source list, retrieval focus, and no-fabrication boundaries.

Organize task-related materials into clear knowledge-base sections and design retrieval questions that help AI find evidence.

A personal knowledge-base section structure, the boundary of each section, and 5 retrieval questions.

Distinguish tasks AI can answer directly from tasks that need external tools, and set human confirmation points for tool actions.

A tool-calling checklist listing task steps, possible tools, what AI may do, and what a human must confirm.

Judge whether a task benefits from an Agent-style workflow and design its goal, steps, sources, tools, confirmation points, and final output.

A lightweight Agent flow sketch including goal, steps, sources or tools, confirmation points, and final output.

Write lightweight quality checks and boundary rules for your AI task so outputs can be checked and behavior can be controlled.

A lightweight AI quality checklist and a set of guardrail rules.

Integrate all previous exercises into an executable, checkable, and improvable AI usage plan.

A complete My AI Usage Plan and an optimization-order table.

课程亮点

Carry one real need through the whole course, from task instruction to final AI usage plan

Place Prompting, RAG, tool calling, Agents, and Evals inside one executable workflow

Produce part of the plan in every lesson, then assemble a submit-ready and reusable final artifact

Train when to add sources, formats, tools, and guardrails instead of stacking AI buzzwords