The Problem: Communication was scattered across too many platforms, leading to missed messages and overwhelmed players, alongside a need to give players 24/7 self-service access to their own film and stats.

Process:

• Mapped existing communication channels and interviewed players and staff on habits and frustrations with current tools
• Benchmarked other sports team communication apps, then prototyped and tested extensively with players and staff

Outcome: Adopted across both team levels, consolidating communication onto fewer platforms and giving players direct, self-service access to team-provided film and stats for the first time.

The Problem: Agents could reproduce established patterns inconsistently or improvise poor design decisions, while overly rigid rules risked suppressing useful experimentation.

Process:

• Tested agent-generated interfaces repeatedly, identifying failures across components, responsiveness, accessibility, visual language, and interaction patterns
• Released the system to independent testers, using divergent prompting styles to uncover failures I had learned to anticipate

Outcome: Generated interfaces now consistently use centralized components and variables while remaining capable of creating novel patterns when needed.

The Problem: The existing filter system required extensive onboarding to use effectively, creating a real barrier for players and new staff.

Process:

• Interviewed and surveyed staff across experience levels to map comfort with the existing tool
• Benchmarked other LLM-driven search patterns for query suggestions, filter detection, and error handling

Outcome: Cut typical search time roughly in half and let players query the database independently for the first time.

The Problem: Comparisons happened manually across separate dashboards, making it hard to spot key differences, surface correlations, or track a team's change over time.

Process:

• Observed users narrating their process through the existing two-tab workaround, then gathered their expectations for an ideal view
• Benchmarked comparison patterns in other sports tools, then simplified through repeated testing to remove intimidation factor

Outcome: Adopted as the go-to comparison tool across the user base, with users reporting significant time savings.

The Problem: Inconsistent components, styling, and undocumented usage patterns created unnecessary design decisions and made implementation less predictable.

Process:

• Audited existing patterns and usage to consolidate unnecessary variations and identify foundational inconsistencies
• Expanded and documented reusable components, variables, and interaction guidelines around the ecosystem’s recurring needs

Outcome: The expanded library now covers most recurring needs, accelerating new-page production while maintaining consistent visual and interaction patterns.