AIUnstuck by LuchaFit · AI product design, privacy
AI coaches can trust with athletes’ private thoughts
Teams finishing LuchaFit’s Unstuck mental performance course wanted to keep going, but what athletes write about pressure, weight, and relationships is deeply personal. I designed an app where AI synthesizes each athlete’s logs into guidance, and decided what it may reveal before designing anything else.
- Role
- Product design and build
- Status
- In development, fall 2027
- Platform
- Mobile-first web
- Users
- Athletes, coaches, parents
- Challenge
- Keep teams engaged after the course, using data athletes would never share publicly.
- What I did
- Wrote the PRD, designed the privacy model and the AI synthesis, and am building it with Claude Code.
- Outcome
- An approved PRD and a build underway, as the product that keeps teams after the course.
Switch views to see what each person gets from the same week of journaling. Sample data.
Journal
- Monday
Down 3 lbs since Saturday. Skipped lunch to make weight and felt flat in live goes.
- Wednesday
Got into it with my practice partner. He keeps going hard when I’m cutting. I didn’t say anything.
- Friday
Talked it out with him before practice. Best live goes I’ve had all month.
For you this week
- You made weight the hard way this week. Want to build a cutting plan with your coach?
- Talking it out with your partner worked. Write down what you said, for next time.
- Your confidence climbed all week. Carry Friday into Saturday.
Need to know
- Weight cutting was erratic this week.
- Friction with a practice partner early in the week, resolved by Friday.
- Energy and confidence trending up going into the weekend.
Suggested: check in one-on-one about their weight plan before Saturday.
Need to knowRead-only, same as coach view
- Weight cutting was erratic this week.
- Friction with a practice partner early in the week, resolved by Friday.
- Energy and confidence trending up going into the weekend.
The business problem
The Unstuck Wrestler is a season-long mental performance course for teams. Many teams that took it didn’t want to retake the same course the next season, but they still wanted to be customers. Without something new, the relationship would end after one year.
The answer had to keep athletes practicing what they learned, give coaches a reason to keep paying, and grow with the athlete over several seasons.
What season one taught us
The course’s first season gave us something better than interviews: we watched what athletes actually did.
- They won’t share in publicIn a community that includes direct competitors, high school and college athletes wouldn’t post their scores, values, weaknesses, or almost anything with their name on it.
- They will share in privateParticipation worked because homework went to LuchaFit confidentially. Athletes were honest when they knew who would read it.
That shaped the core principle: athletes write privately, and AI turns it into a short, need-to-know signal for the adults around them.
Designing the permission model
Before designing screens, I decided who sees what. This table became the backbone of the PRD.
| Athlete | Coach | Parent | |
|---|---|---|---|
| Journal entries | Full access | Never | Never |
| AI summary | Personal guidance | Need-to-know synopsis | Same as coach, read-only |
| Recommendations | For the athlete | For the coach | Same as coach |
| Red-flag alerts | One disclosure at onboarding | Head coach only | No |
| Control | Owns their data; 18+ can turn off parent access | Team view | Follows the athlete’s setting |
The summary never includes overly personal details, such as menstrual health or the specifics of private conversations, even when they appear in a journal entry. Try switching views in the demo at the top of this page.
Key decisions
Tell athletes each time a coach is alerted
One clear, general disclosure at onboarding, and no notice on individual alerts
WhyPer-alert notices would teach athletes to stop writing about hard things. Disclosing the rule up front keeps them informed and keeps the journal honest.
Give parents their own view and summary
Parents see exactly what the coach sees, read-only
WhyTwo different summaries of the same week invite two different stories. One shared view keeps adults aligned.
Track weight as numbers only
Log weight as numbers and as cutting behaviors
WhyBehaviors like skipped meals predict trouble before the scale does, and they’re what the red-flag logic needs.
Fold it into a larger team operations app
Keep Unstuck a standalone product with a flat annual team license
WhyA focused product is easier to sell to teams that already trust the course, and easier to price.
Designing the daily habit
None of this works unless athletes log consistently. I designed for the 30 seconds after a hard practice:
- Voice logging, so an athlete can talk instead of type.
- Daily reminders and seasonal phases with fresh content, so the app doesn’t go stale by February.
- Course tools inside the app: reset routines, competition plans, and visualization.
- The Unstuck assessment retaken periodically, so progress is visible over a season.
Closing the loop for coaches
- AI links mental habits to competition results, so coaches see what actually moves performance.
- Weekly team-talk prompts turn individual trends into something a coach can say to the room.
- Athletes can share wins with their coach on their own terms.
Product strategy
The app is the first step on a ladder that grows with a team over several seasons:
- 01Year 1The Unstuck course
- 02Year 2+The app
- 03Add onApp plus a second-year course
- 04PremiumApp plus live sessions with the coach
Athletes own their data and can take it to a solo plan, so the relationship can continue even after an athlete leaves the team.
Where it stands
The PRD was approved in September 2026, and I’m building it now with Claude Code, targeting the fall 2027 season. The first things I’ll validate with real teams are whether athletes log on their own after week two, and whether coaches act on the need-to-know summaries.