A leading European sports league: an AI & automation framework for internal operations

The league wanted to modernise internal workflows and communication using AI and automation, without adding friction for non-technical teams. Multiple internal teams relied on manual, time-consuming processes for communication, meeting documentation, reporting and publishing — processes that were inconsistent across teams, hard to scale, and dependent on individual effort rather than a system.

Despite the organisation’s digital maturity, meetings still produced unstructured notes, reports needed repeated manual formatting, and internal updates went out as long documents nobody read. The real question wasn’t “how do we add AI?” It was “how do we introduce automation in a way people trust?”

• Led UX design and system thinking across multiple AI-driven internal tools
• Mapped existing workflows to find where automation could reduce manual effort without removing human oversight
• Designed interfaces and interaction flows for AI-assisted tools aimed at non-technical internal users
• Worked closely with engineering and stakeholders to align design decisions with feasibility and business priorities

Before designing any single tool, I mapped how information and tasks moved across teams — where duplication existed, where handoffs caused delay, and where automation could realistically reduce effort without removing necessary human judgement. That mapping held to four principles across all four tools:

1. Automate the repetitive, not the judgement — tools removed manual, repetitive effort while keeping humans in control of decisions that needed context or nuance
2. Design for non-technical users — complexity was hidden behind simple, guided interactions so any internal team could use the tools without a technical background
3. Make outputs trustworthy, not just fast — review and correction steps were built in from the start, not blind automation
4. Fit into existing workflows — tools slotted into how teams already worked, rather than asking them to change everything at once

Internal updates and knowledge-sharing relied on manual meetings and written summaries — time-consuming to produce and inconsistent in quality and reach.

I designed an AI-driven video podcast tool that turned internal updates into a consistent, easily consumable video and audio format, scripted via GPT-4, voiced with ElevenLabs and assembled with branded motion templates. It replaced the manual work of producing internal communications with something teams actually watched.

The result: a roughly 70% reduction in internal communication workload, alongside higher engagement with internal updates and less manual production time.

Meeting notes and follow-ups were inconsistent, often incomplete, and depended on whoever happened to be taking notes — context got lost and accountability for follow-up action was unclear.

I designed an AI meeting intelligence agent that transcribes meetings via the Whisper API, then uses GPT-4 to extract actions and summaries and auto-build the follow-up documentation — capturing and structuring meeting content automatically instead of relying on manual note-taking.

The result: around 4 hours saved per week per team, more consistent documentation, and clearer accountability for what happens next.

Reporting across teams was manual, slow and inconsistent in format, making it hard for leadership to get timely, comparable insight.

I designed a data visualisation and reporting framework that unified match, sponsorship and audience metrics and automated the assembly and presentation of key data, standardising how reports looked and were delivered.

The result: near real-time executive reporting, with consistent branding and format across reports and faster access to insight for decisions.

Publishing content — documentation, updates, internal or external materials — needed manual formatting and design work for every single piece, slowing publication down and creating repetitive work for design teams.

I designed a markdown-to-publication engine, rendered to branded PDFs and microsites via Pandoc, that let non-technical staff produce polished, on-brand output straight from a template — removing the repetitive manual formatting behind every piece.

The result: simplified documentation workflows and less repetitive design work, freeing design teams for higher-value work.