Beyond Prompts: From AI User to System Designer
My friend was recently diagnosed with ADHD. It was a profound moment for her—finally understanding why those years of struggling with focus and motivation weren't character flaws but simply her brain working differently. Yet alongside this relief came overwhelming uncertainty: figuring out medications, creating new routines, and adapting her life to work with her unique neurological wiring rather than against it.
I wanted to help, but not with generic advice or well-meaning platitudes. I wanted to create something genuinely useful—something that could serve as an on-demand guide tailored specifically to her situation. So I built her an AI ADHD coach that combined the scientific insights of Dr. Andrew Huberman with my productivity and life design strategies.
For me this project represented something I've believed since my days at Google: technology should serve people, not the other way around. Yet our current trajectory often seems reversed—our attention and desires are increasingly commoditized, harvested, and manipulated by the very tools meant to empower and connect us.
And since the advent of generative AI, there's been a concerning obsession with increased productivity and output, fueled by greed and fear-based rhetoric about automation and job replacement. It seems like no one's really thrilled about this trajectory, yet everyone's being dragged behind a runaway train. Not even the mega tech companies or governments can stop the momentum because of the arms race nature of AI development.
That's why I wanted to provide not just a different perspective, but a practical demonstration of what's possible when we set aside the competition lens to embrace service and creativity.
The Problem with Conventional AI Approaches
The second area I wanted to address with this video was how most current approaches to AI fall into predictable patterns that limit their potential impact:
Task-oriented prompting: Treating AI as a request-response system for isolated tasks
Content generation focus: Prioritizing output volume over meaningful transformation
Tool-first thinking: Starting with AI capabilities rather than human needs
Generic instructions: Providing minimal context and relying on AI's general knowledge
Passive consumption: Accepting whatever the AI produces without strategic direction
Tactical application: Using AI for short-term gains without a cohesive strategy
These approaches create a transactional relationship with technology that mirrors larger problems in our digital landscape—consumers rather than creators, directed rather than directing, Automatons rather than autonomous beings.
The following are a several shifts that I wanted my audience to make as they watched the video.
#1 The Why-What-How Framework for Intentional AI Design
To create technology that serves deeper purposes, we need to shift from being prompt engineers to system designers. This requires a fundamental reorientation in how we approach AI, starting with the Why-What-How Framework that I developed through my consulting work:
Why: Begin with Intention
Conventional approach: Jumping straight to what the AI can do
System design approach: Clarifying your purpose and who you're serving
For the ADHD coach, my "why" was helping my friend transform her unique brain wiring into a strength rather than a limitation. This intention guided every subsequent decision.
What: Co-Create the Outputs
Conventional approach: Accepting whatever the AI generates
System design approach: Working with AI to define ideal outputs that fulfill your intention
I worked with ChatGPT to define exactly what kinds of responses would be most helpful for someone with ADHD—empathetic acknowledgment, clear explanations, and limited actionable steps to avoid overwhelm.
How: Let AI Handle Implementation
Conventional approach: Micromanaging every detail of execution
System design approach: Setting guardrails but leveraging AI's capabilities for implementation
Once the intention and outputs were clear, I could let the AI handle the details of writing instructions and crafting responses, knowing the system was designed to serve the core purpose.
This framework shows the major shift with AI. Previously, humans were responsible for all three elements, with the majority of our attention and time dedicated to the HOW. Now we can focus our energy on the WHY. This “pyramid” is now flipped on its head, the very definition of a revolution.
#2 Human-Centric Design Principles for AI Systems
At the core of effective AI system design are human-centric principles that prioritize the needs, experiences, outcome of the people using these systems. Here are a few of these principles demonstrated in the video:
Start with Human Needs (aka Start with the Problem)
The ADHD coach began with understanding my friend's specific challenges—overwhelming information, difficulty maintaining focus, and needing practical routines. The technology served these needs rather than showcasing AI capabilities.
User's Level of Awareness
I specifically instructed the AI to communicate at "the user's level of awareness" rather than using clinical language. This meant casual, real-life concerns rather than technical descriptions, and empathetic responses rather than detached analysis.
Incremental Engagement
For someone with ADHD, information overload is a particular challenge. The system was designed to provide bite-sized, actionable guidance—offering 1-3 specific steps rather than comprehensive plans, and including follow-up questions to continue the conversation naturally.
Blend Expertise
In the words of Steve Jobs, creativity is just connecting things. The power of ADHD coach came from combining multiple knowledge domains—neurological research from Huberman with my practical productivity frameworks. This integration created something more valuable than either domain could provide separately.
Human Agency
Throughout the design process, I emphasized that the system should suggest and support rather than dictate. Users maintain decision-making power, with the AI serving as a guide rather than an authority.
#3 From Human-AI Interaction to Human-AI Integration
What makes this approach fundamentally different is the shift from interaction to integration. Rather than treating AI as a separate entity we communicate with, we design systems that function as extensions of our intentions.
This represents an evolution through three distinct phases:
AI as Tool (or Intern): Using AI for discrete tasks with minimal context (conventional prompting).
AI as Extension (or Partner): Designing systems that embody our intentions and expertise to transform them into collaborators and thinking partners.
The second phase—AI as extension—is where the most potential lies. It allows us to extend our impact beyond the limitations of our time and attention, creating systems that authentically represent our unique perspectives and capabilities.
#4 Evolution from Tool User to System Designer
This approach represents a significant evolution in our relationship with technology. As tool users, we adapt ourselves to technology, learning its commands and accommodating its limitations. As system designers, we shape technology to align with our deeper intentions and values. Technology becomes an expression of who we are.
This shift is particularly meaningful for those of us who span multiple domains of expertise and want to combine them to create greater impact.
The journey from AI user to system designer begins with a single step—creating your first intentionally designed AI system:
Start with purpose: Identify a genuine need in your life or someone else's
Map the knowledge: What unique expertise or perspective can you contribute?
Design the experience: How should this system engage with its users?
Build incrementally: Create a minimal version first, then expand based on feedback
Reflect and refine: Continuously improve based on real-world application
The beauty of this approach is that it doesn't require extensive technical knowledge—just clarity about your intention and willingness to learn through experimentation.
To see the complete process of building the ADHD coach and learn the specific techniques I used, watch the full video below. There, I demonstrate each step of the design and implementation process, including the frameworks and tools that make this approach accessible to anyone regardless of their technical background.
What Will You Build?
As AI capabilities continue to advance at breathtaking speed, the question isn't whether these tools will impact our lives—they already are—but whether that impact will align with our deeper values and intentions.
By evolving from user to system designer, you can ensure that your relationship with AI reflects your authentic self rather than redirecting it. You can create systems that extend your unique perspective and expertise into the world in ways that would otherwise be impossible.
In doing so, you’ll develop a fundamentally different relationship with technology, one where:
You define the purpose, not the algorithm
You shape the interaction, not the platform
You determine the outcome, not the business model
So I ask you: What system will you design? What unique combination of knowledge, values, and perspective do you have that could be transformed into an AI extension of yourself? How might you create technology that truly serves people—yourself included—rather than the other way around?
To see exactly how I built the ADHD coach step-by-step, including the specific frameworks and techniques I used, watch the full video tutorial here. And if you’re interested in enhancing your learning, I’m hosting a Deep Dive Discussion in the Mind Map Nation community this Friday that explores these principles in greater detail plus a live demo of how I’d design an AI assistant from scratch. See details below
🧠 Deep Dive Discussion: Building AI Coaches & Assistants in ChatGPT
Event Details:
Date: April 4, 2025
Time: 8:00 AM Pacific Time
Location: Zoom (RSVP to see Zoom link)
Prerequisites: Watching “I Built an AI Coach for My ADHD Friend in 30 Minutes” video
Join me for our Deep Dive Discussion to transform my YouTube videos into interactive learning experiences where we can collectively explore concepts at a deeper level.
In this event, we’ll use my AI tutorial video as the jumping off point to explore topics like:
Design AI systems with clear purpose and intention
Turn specialized knowledge into actionable guidance
Create personalized systems that grow more valuable over time
Leverage AI as both a thinking partner and execution assistant
As a bonus, I’ll be using one of our community members’ recent project of starting a new bio-art podcast to real time demonstrate how I would create an AI assistant to help with launch strategy and build execution
If you can’t make it live, this event will be recorded and summarized in our Community Library afterwards.
Make the most of your Epiphanies
Thanks for reading, my friend. If you found it useful, you can get a new Epiphany in your inbox every month.
