Insight Work
A practical space for learning AI, agents, prompting, and better ways of working.
Insight Work is a plain-language learning space built to explain how modern tools work, where they fail, and how teams can use them responsibly. Read a guide, take a quiz, try a workshop, or use a playbook to apply the ideas to real work.
Choose a track
Choose your path.
Insight Work is more than AI theory. Pick the track that matches what you want to do. Understand the tools, practice building with them, or improve how the work itself happens.
AI Foundations
Understand AI, prompting, agents, hallucinations, context, and responsible use. The plain-language groundwork everything else builds on.
AI for Builders
Use AI while you program: Claude, coding agents, IDE extensions, MCP, documentation workflows, debugging, refactoring, testing, and planning.
Better Ways of Working
Improve workflows, decisions, documentation, process clarity, business agility, and team alignment, with or without AI in the loop.
Start here
Start with the foundations.
New to this? Follow the steps in order. Each one builds on the last, from how AI works to putting it to use. This is the recommended starting path before you pick a deeper track.
- 01
Understand what AI is, and is not
Start with how a language model actually produces text. No magic, just probabilities and patterns.
- 02
Learn to prompt effectively
Once you know how the model works, learn to ask for what you want and get a reliable result.
- 03
See what happens when AI takes action
Give a model tools and a goal and it becomes an agent. Understand the loop, and the new risks.
- 04
Test your understanding
Each guide ships with a short companion quiz. Use them to check what actually stuck.
- 05
Apply it to real work
From here the path branches: use AI at work, use AI while programming, improve a workflow, run a workshop, or pick up a template. Workshops and playbooks turn the ideas into something you can run on a real task.
Field guides
The core learning path.
Long-form visual guides that explain a big concept in plain language. These three connect into one path, read in order or jump to the one you need.
- 01
AI Is Not Intelligent
A plain-language look at what large language models actually do, including probabilities, tokens, hallucinations, context windows, and pattern prediction, while exploring why so many people mistake the output for genuine understanding.
- 02
The Art & Science of Prompting AI
A practical walkthrough of how to prompt large language models well: the structure, context, and iteration that separate a vague request from a reliable result.
- 03
When AI Takes Action
A plain-language look at AI agents. It covers what an agent really is once you give a language model tools and a goal, how the perceive-decide-act-observe loop works, where memory and planning come from, why agents fail differently than chatbots, and the security risks of letting AI act.
Quizzes & challenges
Check what actually stuck.
Short interactive quizzes that go with each guide. A few minutes to see whether the ideas landed, no sign-up, no score kept.
AI for Builders
Using AI while building software, without losing control of the work.
These workshops focus on how to use AI as a programming partner: planning changes, writing better prompts, giving the right context, reviewing generated code, using IDE tools, working with MCP, and validating results through tests, documentation, and version control.
Claude for Developers
Use Claude for planning, debugging, refactoring, documentation, and code review without blindly accepting its output.
Context Engineering for Code
Give an AI the right files, constraints, examples, goals, and acceptance criteria before asking it to write or change code.
AI Coding Workflow
A practical workflow for moving from idea to implementation: plan, scope, prompt, generate, review, test, commit.
MCP and Tool-Connected AI
A beginner-friendly look at what MCP is, why tool access matters, and how connected tools change the way AI helps with development.
IDE Extensions and Marketplaces
Evaluate coding extensions and marketplace tools without installing everything randomly. Focus on usefulness, safety, permissions, and workflow fit.
Debugging with AI
Use AI to understand errors, inspect logs, form hypotheses, and test fixes instead of just asking it to fix this.
Refactoring with AI
Improve existing code safely by working in small steps, keeping behavior stable, and reviewing every change.
Testing AI-Generated Code
Verify generated code using unit tests, manual testing, edge cases, and clear acceptance criteria.
Building Small Apps with AI
Build a small feature or app with AI while staying in control of architecture, quality, and scope.
AI for Builders · Guides
AI in the Developer Workflow
Practical guides for using AI inside your actual development tools. Learn how to work with AI in VS Code, Cursor, Copilot, Claude, and Claude Code, while setting the right context, reviewing changes safely, debugging issues, refactoring carefully, and extending assistants with skills, plugins, and reusable workflows.
Using AI in VS Code
Use AI assistants inside VS Code through extensions, chat, inline edits, autocomplete, and terminal support, without leaving your editor.
Using Cursor
Cursor-specific workflows: project context, composer and agent behavior, reviewing changes, and staying in control of the edits.
Using GitHub Copilot
Copilot chat, autocomplete, inline edits, and PR help, plus when Copilot is genuinely useful and when it is not enough on its own.
Using Claude for Development
How Claude helps with planning, debugging, refactoring, documentation, and code review, used as a partner rather than an autopilot.
Using Claude Code
Terminal-based, agentic coding workflows: safe permissions, reviewing file changes, and scoping a task so the agent stays on track.
Skills, Plugins, and Extensions
Extend AI tools with Claude Skills, plugins, MCP servers, custom instructions, project rules, and reusable prompts that fit your workflow.
Giving AI the Right Context
Provide the files, errors, requirements, architecture notes, conventions, expected behavior, and constraints an assistant needs to be useful.
Planning Before Coding
Ask AI for a plan before implementation: scope, architecture, impacted files, risks, edge cases, and tests, so the work starts on solid ground.
Reviewing AI Code Safely
Inspect diffs, challenge decisions, avoid unnecessary complexity, check dependencies, validate behavior, and keep commits small.
Debugging with AI
Use AI to understand errors, form hypotheses, inspect logs, isolate causes, and test fixes, instead of pasting an error and hoping.
Refactoring with AI
Improve code structure while preserving behavior: work in small steps, update tests, and validate the final diff before you commit.
Workshops
Hands-on, applied to real work.
Guided exercises that take a concept off the slide and onto a real task you can work through.
Insight Work is not about collecting AI tips. Each workshop is built to help you practice a real skill, apply it to a realistic task, and leave with something reusable: a prompt, checklist, workflow, template, or working example.
AI Use Case Workshop
Map a real problem, decide if AI is actually useful, define the expected value, risks, data needs, and human review points.
Prompting for Real Work
Practice turning vague requests into clear prompts with context, constraints, examples, expected output, and review criteria.
Workflow Friction Finder
Analyze a real workflow to find bottlenecks, unclear handoffs, repeated decisions, delays, and automation opportunities.
From Documentation to Process Flow
Turn messy documentation into a clear step-by-step process flow that teams can understand, validate, and improve.
Decision Clarity Workshop
Clarify who decides, what information is needed, what options exist, what risks matter, and when a decision must be made.
AI-Assisted Meeting Prep
Use AI to prepare for meetings by summarizing context, identifying open questions, drafting agendas, and surfacing risks.
Change Readiness Workshop
Prepare a team for a new tool, process, or workflow by mapping what needs to be documented, trained, supported, and measured.
Experiment Design Workshop
Turn an idea into a small experiment with a hypothesis, success signal, timeline, risk, and next step.
Playbooks & templates
Reusable tools you can keep.
Practical canvases and checklists you can lift straight into your own work.
AI Use Case Canvas
Map a candidate use case end to end before you build, so you know whether AI is even the right tool.
Prompt Evaluation Checklist
Score a prompt for clarity, context, and constraints before you trust what it returns.
Agent Readiness Checklist
Decide whether a task is safe to hand to an agent, and what guardrails it needs first.
Change Readiness Checklist
Gauge whether a team is set up to adopt a new tool or a new way of working.
AI Coding Session Brief
Define the goal, files, constraints, risks, and expected output before you ask AI to write any code.
Prompt Review Checklist
Verify a prompt includes enough context, constraints, examples, and success criteria before you run it.
MCP Readiness Checklist
Decide whether a tool or workflow is safe and useful to connect to an AI assistant.
Debugging Notes Template
Track symptoms, hypotheses, tests, findings, and fixes in one simple, repeatable format.
Better ways of working
Where the AI learning meets real work.
Material on business agility, workflow clarity, decision-making under uncertainty, running experiments, and keeping a team aligned. This is the bridge between learning the tools and improving how the work actually happens.
Process Flow Basics
Map work clearly so people can see the steps, decisions, handoffs, exceptions, and pain points in a process.
Better Documentation
Write documentation that helps people do the work instead of only describing the work.
Outcome Over Output
Shift from "build this" to "what problem are we solving, for whom, and how will we know it worked?"
Powerful Questions
A practical set of questions for business analysis, stakeholder alignment, discovery, problem framing, and decision-making.
Stakeholder Alignment Canvas
Map stakeholders, expectations, concerns, decision rights, dependencies, and communication needs before work begins.
Workflow Ownership
Clarify who owns a process, who supports it, who approves changes, and who is responsible when something breaks.
Feedback Loops
Design simple feedback loops so teams can learn from users, errors, delays, escalations, and operational signals.
Handoff Quality
Improve handoffs between teams by defining what must be complete, clear, validated, and traceable before work moves forward.
Prioritization Without Chaos
Use simple criteria to compare requests, reduce noise, explain tradeoffs, and make prioritization more transparent.
Adoption After Launch
Plan what happens after a process, tool, or change goes live: support, training, measurement, feedback, and continuous improvement.
