Your First AI Prompts as a Fitness Coach
The difference between AI output that wastes your time and output you can actually send to a client is almost always the prompt. A vague prompt produces a vague program. A specific, structured prompt produces a draft you can edit in three minutes and send.
What You'll Learn
- The CLIENT framework for fitness-specific prompts
- Five copy-paste prompt templates you'll use weekly
- How to provide context so AI generates safe, appropriate programs
- Common prompt mistakes that lead to bad coaching output
The CLIENT Prompt Framework
Use the CLIENT framework whenever you're writing a prompt for a client-facing task. It maps directly to the information you'd gather in any decent intake.
- C — Client profile: Age, sex, training age, injury history, lifestyle constraints
- L — Level and goal: Beginner/intermediate/advanced, goal (hypertrophy, fat loss, performance)
- I — Inputs available: Equipment, training days per week, session length
- E — Expectations: Output format (4-week block, weekly meal plan, check-in reply)
- N — Non-negotiables: Constraints, contraindications, scope of practice limits
- T — Tone: Coaching voice (firm, friendly, evidence-led, hype)
A good prompt hits all six. A weak one says "make me a workout."
Example: Weak Prompt vs. CLIENT Prompt
Weak:
Write a workout for a client who wants to build muscle.
CLIENT-style:
You are an evidence-based hypertrophy coach. Write a 4-day upper/lower split for an intermediate client.
- Client: 34-year-old woman, 4 years training experience, history of right shoulder impingement (avoid behind-the-neck pressing and upright rows)
- Goal: Hypertrophy with emphasis on glutes and back
- Equipment: Commercial gym, full access
- Sessions: 4 days/week, 60 minutes max
- Format: Markdown table per day with exercise, sets, reps, RIR, and a 1-line cue
- Constraints: Stay within 16-20 working sets per session, include at least one unilateral lower-body movement per lower day
- Tone: Concise, professional, no fluff
The second prompt produces a usable draft. The first produces a generic program you'll throw out.
Five Copy-Paste Prompt Templates
Template 1: Program Block Draft
You are a [strength/hypertrophy/endurance] coach. Write a [X]-week mesocycle for a [level] client.
Client: [age, sex, training history, injuries] Goal: [primary goal + any secondary goal] Equipment: [what they have] Sessions per week: [N], duration: [N min]
Format: Markdown tables, one per session, with exercise / sets / reps / load (% 1RM or RPE) / rest. Include warm-up and cooldown. Add a short progression rule for the block.
Template 2: Weekly Check-in Reply
You are my coaching assistant. Draft a check-in reply to my client based on the message below. Keep it under 200 words. Use a warm but accountable tone.
Acknowledge wins, address one priority issue, give one specific action for next week. Do not give medical advice.
Client message: [paste message] Client context: [stats, goal, current phase, anything I told you previously]
Template 3: Meal Plan Options
Generate 3 sample days of meals for a client targeting [calories] kcal and [protein] g protein. Carbs and fats can flex.
Constraints: [dietary preferences, allergies, dislikes, budget level] Style: [Mediterranean / high-protein / vegetarian / etc.] Output: Per day, list 4 meals (breakfast, lunch, dinner, snack) with food items, portions in grams, and an estimated macro line. Add a 5-item grocery list at the end.
Template 4: Social Media Caption
Write 5 Instagram caption variations for a Reel about [topic, e.g., "why your bench press has stalled"]. My niche is [niche]. My voice is [voice].
Each caption: hook line, 3 short body lines, CTA. Include 8 niche hashtags at the end. No emojis unless they earn their spot.
Template 5: FAQ Answer for Onboarding Doc
Write a 120-word answer to this client question for my onboarding handbook: "[question]". My coaching philosophy is [1-2 lines]. Tone: clear, plain English, no jargon. End with one practical action the client can take this week.
Why Context Beats Cleverness
The biggest mistake new AI users make is treating the AI like Google. They type a short query and hope. The AI cannot read your client's intake form, see their last training block, or remember that they hated lunges. You have to tell it.
Here is a useful habit: keep a short "client context block" saved as a snippet for every active client. Something like:
Client: Maria, 41 F, intermediate, 3x/week, 60 min, home gym (DBs to 30kg, bench, pull-up bar). Goal: fat loss + maintain strength. Knee pain on deep squats. Hates burpees. Prefers concise messages, no exclamation points.
Paste that block at the top of any prompt about Maria. Quality goes up immediately.
Common Prompt Mistakes to Avoid
- No client context — AI invents a generic profile
- No equipment list — AI assumes a commercial gym every time
- No format spec — You get prose when you wanted a table
- No scope limits — AI cheerfully writes medical or supplement claims
- Asking for too much at once — A 12-week program plus nutrition plus messaging in one prompt is a recipe for shallow output
- Not iterating — First draft is rarely the final draft. Reply with "Replace day 3 with a pull-focused session" instead of starting over
Try It Yourself
Open ChatGPT, Claude, or Gemini and try this exact prompt. Then iterate twice with follow-up requests:
You are a personal trainer. Write a 3-day full-body program for a 52-year-old male beginner with a desk job, mild low-back stiffness, and a goal of general health and strength. Equipment: dumbbells up to 20kg, a bench, a resistance band. 45 min per session. Output: a markdown table per day with exercise, sets, reps, tempo, and a coaching cue. Add a "first session priorities" section above the tables.
Now ask: "Make day 1 less spine-loaded — replace the bent-over row with a chest-supported variation and the goblet squat with a step-up." See how the AI adjusts. That iteration loop is most of the skill.
Key Takeaways
- The CLIENT framework gets you safe, specific, usable output
- Always specify client profile, equipment, sessions, format, constraints, and tone
- Save a short context block per client and paste it into every prompt about them
- Iterate with follow-up edits instead of rewriting prompts from scratch
- AI gives generic answers to generic prompts — your specificity is the lever

