Original Insight
“I don’t want a prompt that writes a specific article to a specific audience about a specific topic for a specific word count. I want an interactive prompt that works with the user and then infers all of these variables dynamically, and then thinks and then runs.” — Lou
Expanded Synthesis
There is a ceiling that most AI users hit within their first six months of serious prompting — and it is not a ceiling imposed by the technology. It is a ceiling imposed by the way they think about the technology.
Most knowledge entrepreneurs, coaches, and creators approach AI as a task-executor. They type a request, receive an output, and iterate from there. This is conversational prompting, and while it is better than nothing, it is fundamentally labor-intensive: every new task requires a fresh setup, a fresh round of iteration, and fresh time expenditure. The value extracted is real, but it is not compounding.
Lou’s June 5 masterclass introduced a more sophisticated model: what he calls process prompts — a new level in the hierarchy of AI use that transcends the usual prompt library approach.
The hierarchy he outlined looks like this:
- Conversational prompting — freestyle chat with the AI per task
- Specific prompts — reusable instructions for defined use cases
- Super prompts / custom assistants — sequences of operations plus context captured as a system
- Meta prompts — prompts that generate other prompts
- Process prompts — dynamic, interview-driven systems that infer variables and think before executing
The crucial distinction at the highest level is generalization. A process prompt does not execute a fixed task — it creates the architecture to execute any version of that task. It interviews the user, reasons about the request, applies best-practice frameworks, and produces output that is custom to the moment, not hardcoded at the design stage.
The practical implication is significant: instead of building a prompt that writes an email for your audience, you build a system that learns what kind of email to write based on the context the user provides in real time. Instead of a prompt for generating social posts, you build a coach-trained system that can dynamically construct posts for any client, any topic, any goal — just by asking the right questions first.
For high-performing coaches, this reframing is everything. The most expensive thing you do each day is not the client work — it is the setup work: the thinking, the context-setting, the explaining-to-the-AI-what-you-want. Process prompts collapse that setup cost. Once designed, they are operational assets, not recurring tasks.
Lou also demonstrated what he calls “infinite prompting” — a technique borrowed from a Michael Simmons article — where every step of the reasoning process is written out in a separate artifact/canvas. This forces the AI into deeper deliberation per step, producing far richer intermediate outputs that feed a higher-quality final product. The trade-off is time (each session takes 10+ minutes), but the quality differential is pronounced.
The coaching insight here is this: the goal is not to teach the AI what to do — it is to teach the AI how you think about what to do, so that it can think that way dynamically on your behalf. You are not building a script. You are building a judgment model.
This matters deeply for coaches, consultants, and course creators because your competitive advantage is not information — the AI has more information than any human. Your advantage is your specific way of thinking about your domain. If you can codify that into a process prompt, you make that advantage scalable.
The common pitfall is building too specifically. Coaches create a prompt that writes their welcome email, one that creates their intake form, another for their follow-up sequence. Each is useful in isolation. But none compound. None transfer. None get better with use. Process prompts, by contrast, are designed to stay useful as your business evolves, because they hold the logic, not the specific execution.
Practical Application for PowerUp Clients
The 3-Level Prompt Audit
Ask clients to audit their current AI use in three categories:
| Level | Description | Example | Problem |
|---|---|---|---|
| Conversational | One-off requests typed fresh each time | ”Write me a subject line” | No reuse, time-intensive |
| Specific | Saved prompts for recurring tasks | Onboarding email template | Fixed, breaks when context changes |
| Process | Interview-first dynamic systems | ”Describe the client and goal; AI writes the email” | Requires design investment |
The Process Prompt Design Exercise:
- Pick one recurring writing or decision task you do every week for clients or in your business
- Write out the 5 questions you would ask a brilliant intern before asking them to complete that task
- Identify 2-3 best-practice frameworks that inform how you’d evaluate a great output (e.g., for emails: clarity, hook strength, emotional resonance, call to action)
- Write a meta-prompt that asks Claude or ChatGPT: “Use these questions and these frameworks to design an interactive system that handles this task generically”
- Test the system with 3 different client scenarios and refine
Journal prompt for clients: What is one task I do repeatedly where I bring real judgment and expertise — but spend more time on setup than on the actual insight? That is the process prompt opportunity.
Additional Resources
- Michael Simmons’ Infinite Prompting framework — look for the article on Substack or Medium regarding multi-artifact prompting
- Claude’s built-in artifact system for multi-step reasoning
- Insight - Codify Your Judgment Into Skills, Not Just Prompts — closely related; this insight is the tactical implementation of that strategic principle
- Insight - Build Tiny Tools That Remove Real Friction — process prompts are the higher-level version of the “tiny tools” concept
Evolution Across Sessions
Baseline — first instance of this framework. This is Lou introducing his most advanced prompting philosophy to the mastermind group for the first time. The concept of “process prompts” as a category distinct from meta prompts is newly coined terminology in this session.
The June 5 session is notable because Lou is running a hybrid format: a teaching event for prospective members of the AI Leaders Mastermind. The audience is somewhat mixed (experienced and new), but the content is his most sophisticated current thinking on AI leverage. This insight should be treated as a foundational principle for the mastermind’s operational philosophy going forward.
Watch for how this concept evolves in later sessions as members begin implementing their own process prompts and reporting results.
Next Actions
- For me (Lou): Create a “process prompt starter kit” — a guided conversation template that takes any member through the 5-question design exercise above, producing a ready-to-use process prompt for their top recurring task
- For clients: Complete the 3-Level Prompt Audit for their business this week; identify one task to elevate from specific to process level