Program · Individuals
Advanced Prompting for Deep Work
For power users who want AI to help with real thinking: analysis, strategy, and complex creative projects. We'll move beyond one-shot prompts into multi-step systems you can reuse for the hardest parts of your work.
You'll learn how to break big problems into stages, design prompt chains, run longer research sessions, and build a personal prompt library that actually sticks — across tools like ChatGPT and Google Gemini.
Program overview · Audio
A short overview of how this program helps you design, build, and use personal AI assistants without writing a line of code.
What you’ll be able to do after this program
Less “AI as a clever autocomplete,” more structured thinking support for serious work.
- • Decompose complex tasks into stages where AI can help without taking over your judgment.
- • Design multi-step prompt flows that turn messy inputs into structured, useful artifacts.
- • Run longer analysis and research sessions without losing the thread.
- • Work across multiple tools and models while keeping your own mental model stable.
- • Maintain a small, sharp library of advanced prompts and workflows for your deepest work.
Who this is for
- • Heavy AI users who already know the basics
- • Knowledge workers, strategists, analysts, and leaders
- • Creators and builders doing deep, multi-step work (not just quick drafts)
What you’ll work on
- • One or two real deep-work projects
- • Your own advanced prompts and prompt chains
- • A reusable structure for future complex work
Curriculum at a glance
Four modules to design, run, and maintain advanced prompting systems for serious work.
Module 1
Decomposing complex work into AI-friendly steps
- • Spotting which parts of deep work AI can help with
- • Breaking big questions into staged prompts
- • Designing intermediate artifacts (lists, maps, outlines)
Module 2
Multi-step prompting systems
- • Prompt chains vs. single mega-prompts
- • Review, refine, and re-run loops
- • Designing “prompt scripts” you can reuse
Module 3
Long-running research & analysis sessions
- • Structuring multi-session work with AI
- • Working across documents, notes, and tools
- • Capturing reasoning, not just final answers
Module 4
Operating your own deep work prompting library
- • Versioning and documenting advanced prompts
- • Avoiding brittleness across models & tools
- • Building a personal library you actually use
1. Decomposing complex work into AI-friendly steps
Deep work becomes manageable when you can see the steps. AI is much more useful there, too.
We start by mapping the kind of work you want help with: analysis, strategy, writing, planning, research. Then we break it into stages where AI can assist without replacing your judgment.
- • Identifying “thinking moves” in your own work
- • Separating discovery, structuring, and drafting
- • Using intermediate artifacts to avoid overwhelm
- • Designing step-by-step flows you can run repeatedly
Exercise: deep work decomposition
You'll take one real deep-work project (strategy doc, analysis, long-form piece, etc.) and:
- • Map out 5–9 key steps you usually take
- • Mark which steps AI can safely support
- • Draft a first "step sequence" for AI-assisted work
2. Multi-step prompting systems
Stop relying on one giant prompt — build prompt chains and loops you can trust.
Here we design multi-step flows: a sequence of prompts that each do one job, with room for your review in between. Less magic, more control.
- • The trade-offs of mega-prompts vs. prompt chains
- • "Review & refine" loops for higher-stakes work
- • Breaking work into “input → transform → check” steps
- • Writing prompt scripts you can run again later
Your first prompt system
You'll design and test a prompt system (not just a single prompt) for one of your deep-work tasks:
- • A written “prompt script” with 3–7 steps
- • Clear spots where you review and adjust
- • Notes on where AI tends to slip or drift
3. Long-running research & analysis sessions
Stay oriented over hours or days of work — even when tools and tabs change.
Deep work rarely happens in one sitting. We look at how to keep context and direction across multiple sessions, tools, and documents — with AI as a stabilizing partner.
- • Designing “session briefs” for multi-day work
- • Using AI to summarize and re-orient between sessions
- • Working across docs, notes, and chat windows
- • Capturing reasoning, not just conclusions
Deep work session template
You'll create a template you can reuse for any long AI-assisted work session:
- • A one-page “session brief” structure
- • Check-in and wrap-up prompts for each session
- • A simple way to track insights and open questions
4. Operating your deep work prompt library
Treat your prompts like a tiny product: versioned, documented, and lightweight.
Finally, we turn your experiments into a stable library: small, documented, and resilient when models or tools change.
- • Naming and organizing your best prompt systems
- • Writing short “how to use this” notes for each
- • Handling model changes and behavior drift
- • Deciding what to keep, revise, or delete
Your deep work prompt library
You'll leave with a compact, living library that includes:
- • 3–7 advanced prompting systems you actually use
- • Clear documentation and example use cases
- • A simple review cadence to keep it sharp
Format & logistics
A focused, advanced-level program for serious AI power users.
Schedule
- • 3–4 weeks total
- • Weekly live sessions (60–90 minutes)
- • Deep-work blocks between sessions on your projects
Group size
- • Small cohorts: typically 3–10 people
- • All participants are already regular AI users
- • Great for “power user” circles inside companies
What you leave with
- • A set of advanced prompt systems for deep work
- • Templates for long-running sessions and projects
- • A documented prompt library you can grow over time
FAQ: Advanced Prompting for Deep Work
Questions people usually ask before committing to this level of work.
How advanced do I need to be before joining?
You should already be comfortable using tools like ChatGPT or Gemini a few times a week, know basic prompting patterns, and have at least one serious project or problem you want help with. If you're brand new, the Foundations or Productivity programs are a better starting point.
Will we write code or integrate APIs?
This program is focused on prompting systems and workflows, not engineering. We'll design things you can later automate or integrate if you want, but the core work happens in regular AI interfaces and your usual tools (docs, notes, etc.).
Can I use multiple tools and models?
Yes. We'll design patterns that survive model changes and work across tools. If you primarily use one provider at work, we'll help you adapt the same flows there while keeping your mental model stable.
Is this useful if my company is already building AI tools?
Definitely. This program focuses on how *you* think and work with AI. Many people who build AI products find it clarifies how to design better flows and evaluate their own tools from a power-user perspective.
Ready to turn AI into a true partner for deep work?
This program is for people who want AI in the room for their hardest thinking — with structure, not chaos. You'll come away with systems you can rely on for your most important projects.
If you’d like to run this as a private cohort for your “power-user” group or leadership circle, mention that in your note and we'll share options.