Build Your Second Clinical Brain
This prework exists because the live day is too valuable to spend on basic self-description. Your goal is to arrive with enough raw clinical material that the day can begin with discussion, challenge, and build momentum, not background data entry.
Plan for roughly 90 to 120 minutes total, broken across the week. Don't over-edit. Be specific. The more honest and concrete this is, the more useful your Claude artifact will become.
Suggested pacing across 7 days
- Day 1. Section 0. Set up your Claude Project.
- Days 2–3. Section 1. Clinician Context (longest section).
- Day 4. Section 2. Program Reality Check (you'll need your 3 recent programs in front of you).
- Day 5. Section 3. Real Client Case.
- Day 6. Section 4. Honest Reflection.
- Day 7. Section 5 upload, plus the Ready Check.
Two reasons. First, it cuts the time in half. Second, and more important, speaking your answers tends to surface what you actually think, not the cleaned-up version you'd type. Talk it through. Unfiltered is the goal. Every answer should sound like you talking, not like you writing for an audience. Download: wisprflow.ai
What you're walking into
This is not an AI workshop. It is a clinical build lab where the clinical mind comes first and Claude is used to create leverage around that mind.
AI does not replace the clinical mind. It reveals the quality of it.
If your thinking is vague without AI, it will be vague with AI. If your reasoning is sharp, AI can help you deliver that sharpness with more consistency, speed, and leverage. The day is built around that hierarchy: clinical reasoning first, the human relationship second, technology third.
What you'll leave with
A personalized Claude-based programming artifact that understands who you are as a clinician, who you serve, how you select exercises, how you load and progress, how you modify around pain and constraints, how you communicate, and what your standards are. You will run it against a real client case before leaving.
How this workbook works
Type your answers directly into the fields in each section. Your work auto-saves to your browser as you go. Close the tab, come back tomorrow, your progress is still here.
When you finish Sections 1–4, click Download Worksheet in the bottom bar. A PDF saves to your computer, formatted as PRIMED-Clinician-Context_[YourLastName].pdf, ready to upload to your Claude Project (Section 5 walks you through it).
Your answers save automatically to this browser on this device. The save indicator at the top right of the page shows you the last time your work was saved. Watch for it as you type.
Because the save is local to this browser, complete the prework on one device throughout the week. Work you type on your laptop won't appear on your phone, and clearing your browser's site data or using private/incognito mode will wipe saved progress.
If you need to switch devices mid-week, or you're nervous about losing work, download the PDF early using the button in the bottom bar. It captures everything you've written so far. You can return any time to keep editing and re-download.
What to have ready for event day
- Laptop (fully charged) and charger.
- Claude Pro logged in, with your Project set up and populated.
- 3 real programs you wrote in the last 3 months. Bring real ones, not showpiece examples.
- The real client case you selected.
- Any templates, program formats, check-in forms, or client communication examples you use regularly.
- Notebook or paper for non-digital reflection.
PRIMED-Clinician-Context_[YourLastName].pdf, matching what the event team expects on upload.Set Up Your Claude Project
Your Claude Project is the workspace where your artifact will be built. Setting it up now means your context, programs, and client case are loaded and ready when the live day begins, so you can start building immediately instead of doing setup at 8:30 AM.
Confirm Claude Pro
Claude Pro is required for this event. The free tier doesn't give you the Project capacity, larger context window, or knowledge base features you'll need.
- Go to claude.ai and sign in (or sign up).
- If you don't yet have Pro, upgrade in your account settings. Pro is currently $20/month.
- Confirm "Pro" appears next to your name or in your account area.
- Download the Claude desktop app from claude.ai/download and sign in with the same account. We use Claude Desktop (not the browser) for the prework and the event because file uploads, keyboard shortcuts, and the chat surface are more reliable there.
Pin your model and settings
Pro lets you switch between several Claude models, each with different strengths. For the prework and the event, we are standardizing so everyone in the room is working with the same Claude.
- Open a new chat in Claude Desktop. Click the model selector (usually at the top of the chat window or in the message bar) and choose Claude Sonnet 4.6.
- Turn extended thinking OFF. Extended thinking changes how Claude reasons and slows response time. We want a fast, consistent build experience.
- Turn off any tools (web search, file creation, connectors, computer use). Your Project should rely only on what you put in the knowledge base.
Create your Project
- In the left sidebar of Claude Desktop, locate "Projects."
- Click "Create Project" (or the "+" / "New Project" button).
- Name and describe the project using the snippets below. Both are pre-copied for you.
Add your Project instructions
Project instructions tell Claude how to behave inside this workspace. Find the "Set project instructions" or "Custom instructions" area inside your new project. Copy the block below and paste it in, then save.
You'll refine these instructions during the event. This baseline gets you started.
Locate the Project knowledge area (labeled "Files")
Inside your Project, Claude labels the knowledge area as "Files." This is where you upload documents Claude will reference in every conversation inside the project. You don't need to add anything here yet. Section 5 of this worksheet walks you through uploading your completed worksheet when it's ready.
Test it
Once your Project is created and your instructions are saved, start a new chat inside the Project (not a regular chat) and paste the test prompt below. Your knowledge base is empty at this point. The test is just confirming your Project instructions are saved.
If Claude responds in line with your project instructions and acknowledges your clinical responsibility, your Project is set up correctly.
Clinician Context
Generic AI output starts with generic context. If Claude doesn't understand who you are, who you serve, and what your clients actually need, it will default to average programming for average people. None of you work with average abstractions.
Your niche is not marketing fluff. It is clinical context. Aim for 3 to 6 sentences per answer. Long enough to be specific, short enough to be honest. If you find yourself writing "active adults" or "people who want to feel better," push harder.
Postpartum runners, six months to five years out, mostly age 30 to 42, returning to longer mileage and races they used to do pre-baby. Frequent overlap with leakage, prolapse symptoms, low back or hip pain, and the all-purpose "things just don't feel right." Most are juggling young kids, fragmented sleep, and a return-to-work timeline. I'm empowering women on their wellness journey through evidence-based care.
What type of clinician are you?
Your discipline, your training, the work you actually do day-to-day.
Who do you primarily work with?
Be specific. Age range, life stage, sport or activity, common diagnoses or presentations, common life context.
What are the most common goals your clients bring to you?
What are the most common barriers your clients face?
Time, money, fear, prior bad experiences, beliefs about their body, life stress, family demands, work schedule, equipment access, anything else.
What are the most common pain or performance problems you help with?
What makes your client population different from a generic fitness or rehab population?
This is the question that protects your artifact from sounding like every other AI tool. What is true of your people that would not be true of a random gym member or a random rehab patient?
What do your clients usually need MORE of?
Examples: confidence, capacity, clarity, accountability, exposure, intensity, consistency, education, autonomy, structure, permission, challenge. Pick the top two or three for your population and explain.
What do your clients commonly misunderstand about their body, pain, training, or progress?
What do you find yourself explaining repeatedly?
The conversations you have over and over. The reframes that show up week after week. These are gold for the artifact.
What parts of programming currently take the most time or mental energy?
What would a generic AI programming tool likely get WRONG about your clients?
This is one of the most important questions in the prework. Be honest and specific. Examples: It might over-medicalize pain. It might underdose strength. It might over-prioritize corrective exercise. It might miss fear or adherence barriers. It might write a technically appropriate plan the client won't follow. It might sound nothing like you.
What would Claude need to understand about you before you trusted it to support your programming workflow?
Your standards. Your non-negotiables. Your communication style. The things you would never want it to assume, overstate, or decide without you.
Program Reality Check
Your programs are receipts. They show what you actually value when time is limited, the schedule is full, and decisions are being made under real-world constraints. Claude does not only need your ideals. It needs your patterns.
Your philosophy lives in your patterns, not your intentions. Pull 3 programs you've written in the last 3 months (real ones, not showpieces) and scan them for 5 to 10 minutes before answering. Answer from what you actually see, not what you wish you saw.
Lower body skews heavily hinge and split-squat patterns. Conventional and trap-bar deadlifts show up almost weekly; RDLs bridge heavier loading and longer runs. Single-leg work is constant: reverse lunges, step-ups, B-stance RDLs. Bilateral squats are rarer than I would have guessed before I looked. Carries show up more than I planned. I am consistently under-prescribing pulling.
Part A. What your programs actually do
Keep your answers tight. A few lines per box is plenty.
Exercise selection
What lower-body and upper-body primary movements show up most often? Which movement patterns (hinge, squat, lunge, push, pull, carry, trunk) do you reach for most?
Loading and dosing
What set and rep ranges show up most often? How do you prescribe intensity, whether RPE, RIR, percentage, feel, tempo, or something else?
Session and weekly structure
What does your typical session look like, start to finish? What does your typical training week look like?
Modification and progression
How do you typically handle pain or low readiness? How do you decide when to progress, regress, or hold steady?
Communication
What words or phrases show up repeatedly in your notes or messages? Do you tend to explain why, or mostly tell the client what to do?
Part B. What your patterns reveal
Now step back from the receipts.
What shows up more than you expected? What almost never shows up?
What do your programs suggest you actually value?
Where do your programs match your stated clinical identity from Section 1? Where do they contradict it?
Where might you be overcomplicating or underdosing?
A rule is a repeated choice with reasoning behind it. A habit is a repeated behavior that survives because it is familiar. Claude should learn your rules. It should not blindly replicate unexamined habits. Defaults are not the problem. Unexamined defaults are.
Real Client Case
The final sprint of the live event has you generate a Monday-ready next-week programming draft for a real client using the artifact you build. This case is the through-line of your entire day. Choose carefully now so you can move quickly when it counts.
Use initials or a pseudonym only. No identifiable client information anywhere in this section. This file will be uploaded to your Claude Project, so the privacy rule isn't optional.
How to choose
- Pick a client who represents the kind of person you actually want this artifact to help you with on an ongoing basis.
- Do not pick your easiest client. Pick someone where a clearer programming workflow would save you time, reduce friction, or improve your decision-making.
- Pick someone clinically relevant but not so complex that the case derails your day. The test is relative to your own practice. Pick a client who sits in the middle of your typical caseload, not at the hard edge of it.
- Ideally, this is a real decision you need to make in the next one to two weeks.
Client identifier: R.W.
Age, life stage, occupation, family or life context: 35, second pregnancy, 9 months postpartum. Works full-time in tech, two kids under four, partner travels for work two weeks a month. Sleeps in 3 to 4 hour chunks most nights.
What you're trying to decide: Whether to add a second short run this week or hold at one and push the strength session. She's been symptom-free for three weeks but leg-day soreness has been spilling into runs.
Client identifier (initials or pseudonym)
Age, life stage, occupation, family or life context
Why they're working with you
The presenting issue, the goal, the story.
How long they've been with you and where you are in the journey
Current pain, symptoms, or performance status
Training history and current capacity
Their beliefs, fears, or misunderstandings
What they think is going on. What they're afraid of. What they've been told before.
Real-world constraints
Time available per week. Equipment access. Schedule. Life stress. Anything that shapes what is actually possible.
What you're trying to decide about their next 1 to 2 weeks of programming
This is the question your artifact will help you answer on event day. Get clear on it now.
What a generic AI would probably get wrong about this specific client
Honest Reflection
The day will expose where your thinking is clear and where it's automatic. You'll get more out of the event if you arrive having already noticed your own patterns. Honest answers here are for you first.
Default warm-ups. Every client gets the same hip CARs, glute medius isolation, and breath drill, almost regardless of what we're about to do that day. It worked for postpartum runners when I built it three years ago. I'm not sure it works for the broader range of clients I see now. I keep meaning to revisit it and don't.
Where am I currently clear in my programming reasoning?
The parts you can defend. The decisions you make with confidence and can explain.
Where am I currently automatic, running on autopilot without re-examining?
Where do I think I'm overly generic, applying the same approach across clients who probably need different things?
Where do I need sharper decision rules?
The places where you find yourself hesitating, second-guessing, or making it up each time.
What is my current relationship with AI tools in my work?
Skeptical, curious, already using, frustrated, hopeful, somewhere else. Be honest.
What would make this event a clear win for me?
Picture yourself on the drive home. What did you walk out with? What changed?
Upload to Your Project
Your final answers from Sections 1 through 4 only become useful to Claude once they live inside your Project's knowledge base. This is the step that turns your prework into context Claude can actually use on event day.
Download your worksheet
Click "Download Worksheet" in the bottom action bar of this page. A PDF will save to your computer, formatted exactly the way the event team expects on upload, including the standardized filename PRIMED-Clinician-Context_[YourLastName].pdf.
If you didn't fill in your last name at the top of this page, the file will default to PRIMED-Clinician-Context.pdf. Fine to upload, but standardized filenames let us help you faster on event day if something breaks.
Privacy check before you upload
Scan the downloaded PDF once. Confirm: no client full names, no identifiable health information, no employer details that could identify a specific person. Use initials or a pseudonym throughout Section 3.
Upload to your Project's Files area
- Open Claude Desktop. Click your PRIMED: Second Clinical Brain Project in the left sidebar.
- Inside the Project, find the section labeled "Files" (usually on the right side of the project page). This is Claude's name for the Project knowledge area.
- Click "Add Content" (or "+"). Select "Upload from device," or drag and drop the PDF directly.
- Wait for the upload to complete. You'll see your document listed in the Files area.
Confirm Claude can read it (the portrait test)
Open a new chat inside your Project. Click the button below to copy the portrait test prompt, then paste it into Claude.
How to grade the portrait
The portrait passes if all four are true:
- It names at least three specifics from your document (a client type, a phrase you use, a constraint, a non-negotiable). Not generic language about your discipline.
- It names one place where your stated identity and your actual program patterns line up, and one place where they don't.
- It ends with your real client and one question you should be sitting with before event day.
- It reads like a sharp colleague describing you, not like a marketing bio.
If two or more of those are missing, the upload likely didn't land. Check the file in your knowledge base and re-run the prompt. If you still cannot pass the rubric after a second try, message the event team at least 48 hours before the event.
Ready Check
Run through this the night before the event.
- My Claude Pro subscription is active and I am logged in.
- My PRIMED: Second Clinical Brain Project is created with project instructions saved.
- My Clinician Context, Program Patterns, Real Client Case, and Honest Reflection are all in my Project's Files area.
- I ran the portrait test and Claude responded appropriately.
- My laptop is charged and packed with my charger.
- I have my 3 programs and my client case ready to bring.
You are not arriving to learn a prompt. You are arriving to build a working clinical asset that reflects your reasoning. The quality of what you build is set by the quality of what you bring. See you on event day.