• Hard-to-find information

  Fans often had to jump between multiple websites and apps to find simple details like match schedules, league tables or player stats.

• Scattered data

  Key information on fixtures, teams and players was spread across different systems, making it difficult to give fans a single, clear view of what they cared about.

• Limited personalization

  Most experiences were one-size-fits-all, not tailored to a fan’s favorite club, players or the type of information they usually look for.

• Need for real-time insight

  The organization wanted fans to get up-to-the-minute information during live games and clear summaries afterward, without having to search.

• Trust and safety for a global audience

  With such a diverse fan base and high public visibility, any AI-driven service needed to behave responsibly, avoiding harmful content, respecting privacy and handling sensitive topics in a safe and unbiased manner.

• Embedded assistant

  Fans can ask natural questions, such as “When is my team’s next match?”, “Show me today’s league table” or “What happened in yesterday’s game?” and get instant answers within the organization’s own digital channels.

• Team and player insights

  The Copilot provides real-time updates on match results, player performances and rankings, as well as always-current league tables, so fans can see how their club is progressing throughout the season.

• Fixture and schedule details

  Supporters get quick access to upcoming matches, kickoff times and venues through a single, simple question rather than browsing multiple pages.

• Live match and event summaries

  During and after games, the Copilot offers easy-to-read timelines: goals, substitutions, major decisions and key moments, so fans can keep up even if they cannot watch live.

Behind the scenes, the experience is powered by:

• Natural Language Processing, so the Copilot can understand everyday questions and respond in a conversational way
• Real-time data feeds, bringing in live match facts and stats so answers are always current
• GenAI, turning complex data into clear, human-like summaries and explanations

To ensure the Copilot was ready for such a large, diverse audience, HCLTech and Responsible AI specialists ran a comprehensive safety and fairness evaluation before launch:

• Three-phase testing program

  The team designed a multi-stage evaluation that looked at 15 key areas of Responsible AI, including safety, fairness and privacy.

• Over 50,000 test questions

  The Copilot was challenged with more than 55,000 prompts, including difficult and “trick” questions designed to probe how it handled sensitive topics, as well as misleading requests or attempts to prompt it to break the rules.

• Single-turn and multi-turn conversations

  Tests covered both one-off questions and longer back-and-forth exchanges to see whether the system stayed safe and consistent over time.

• Fix–verify loop

  Wherever issues were found, the engineering team refined filters and safeguards, then re-ran the tests to confirm the improvements.

• Boosted fan engagement

  The Copilot helped increase interactions on digital platforms by around 40%, as fans discovered how easy it was to get the information they needed through conversation.

• Much faster answers

  Response times dropped from minutes to seconds. Instead of searching through menus and pages, fans now receive instant, clear replies to their questions.

• Personalized experiences at scale

  The assistant tailors answers based on fan preferences – their favorite team, competition or players – delivering more relevant content and a smoother experience.

• High level of safe behavior

  The Responsible AI evaluation showed that Copilot safely handled over 97% of the prompts tested, even under challenging scenarios.

• Clear focus for ongoing improvement

  The testing confirmed that content filters and safety checks remained stable under pressure and pinpointed bias as the main remaining area for continuous refinement.

• Confidence for large-scale rollout

  The organization now has a strong safety baseline and a proven test approach that it can reuse for future updates and features.

• Greater trust from fans

  Supporters can rely on the Copilot to provide accurate, respectful information, helping to build confidence in the league’s digital channels.

• Simpler journeys and less effort

  Fans no longer need to know where data lives or which page to visit. They ask a question and get an answer – reducing frustration and keeping them closer to the competition they love.

• A safer digital environment

  The thorough pre-launch evaluation means that harmful, misleading or inappropriate content is far less likely to reach users, which is crucial for a global audience that includes younger fans.

• A platform for future innovation

  With strong foundations in place, both technically and from a Responsible AI perspective, the organization can continue to extend the Copilot with new languages, richer storytelling and even more interactive fan experiences.

• designing a conversational fan experience around real supporter journeys
• harnessing Microsoft Azure OpenAI to turn complex football data into simple, natural conversations
• and rigorously testing the system against tens of thousands of scenarios to ensure safety, fairness and reliability