AWS for M&E Blog
Holstein Kiel transforms scouting with OpenAI models on Amazon Bedrock
This blog post was co-authored by Markus Brunnschneider, Head of Game and Tactical Analysis, Squad Planning and Scouting, International Football Institute.
Traditional football clubs generate vast amounts of player data but often struggle to transform it into competitive advantages. Holstein Kiel solved this challenge by implementing an AI-powered scouting platform using OpenAI models on Amazon Bedrock, a comprehensive, secure, and flexible platform for building generative AI applications and agents.
We will demonstrate how sports organizations can use OpenAI models on Amazon Bedrock to transform player data into competitive advantages. We will illustrate how Holstein Kiel broke down data silos and created an AI-powered scouting platform that enables better transfer decisions and improved team performance.
If you’re working with sports analytics or managing data silos in your organization, this user case study shows how Amazon Web Services (AWS) can help you extract actionable insights from complex data sets.
Challenge: Eliminate data silos
Holstein Kiel, a traditional German football club, faced data siloing. The club needed to unify large amounts of digital data from multiple sources. These sources included event data from matches, tracking data, and personal training data.
For Holstein Kiel, the data journey meant far more than only introducing new technologies. It needed to be a comprehensive transformation affecting different areas and answering various questions:
- How will departments work together in the future?
- How can decades of experience be meaningfully combined with modern technology?
- How do you measure the success of new approaches?
Therefore, the club wanted to create an innovative solution. This solution would help increase both individual player performance and overall team performance. It would also support better informed transfer decisions.
“Data is an important component for us in evaluating player performance more objectively and identifying development potential early on. It provides valuable foundations for scouting and squad planning—always in combination with the expertise of our coaches and scouts. The crucial factor is generating valuable information from the multitude of individual data points that we can consider in our decision-making,” said Olaf Rebbe, Holstein Kiel Sporting Director & Vice President.
Identifying data sources: Strategic approach
In modern football, there are countless data sources, from merchandising to broadcasting to social media. However, the approach Holstein Kiel uses focuses on data that directly influences sporting performance.
With this focus, Holstein Kiel identified six central areas:
- Workload management
- Mental wellbeing
- Medical care
- Team management
- Scouting
- Match analysis
Scouting and player development are particularly central, as this is where the greatest potential for sustainable competitive advantages lies from both sporting and economic perspectives.
All these areas have one thing in common: improvements directly impact the core business. Every small optimization can make the difference between victory and defeat. This is precisely why the data journey of Holstein Kiel is so crucial—it can make a significant contribution to ensuring the club’s sustainable success.
Head Coach Marcel Rapp, “We have enormous expertise within our team. The better we can deploy and utilize this knowledge, the more successful we will be. Data helps us above all to generate important, objective information in a short amount of time. Making the right decisions quickly is a decisive factor in our daily work.”
Player-centric data architecture: The foundation
The six identified application areas can be applied to additional professional teams and to youth teams, partially to external matches. The solution creates seamless documentation of player development from youth teams to the first team, capturing both sporting development and the economic value of each player throughout their entire club career.
The various data sources are not viewed in isolation. Instead, information is intelligently linked, both interdisciplinary and over longer periods. For this multidimensional connection, a unique key is needed that remains constant across all areas and over time. In football, that’s the players themselves.
This player-centric approach enables Holstein Kiel to create detailed performance profiles that go far beyond conventional scouting methods. Through continuous data collection comprehensive digital twins of players emerge—valuable assets for both internal player development and external transfer options.
The solution space of Holstein Kiel extends across three dimensions:
- Application areas (Topic of Opportunity)
- Teams (Environment of Opportunity)
- Data sources
Everything is held together by the player ID as the central identifier. Figure 1 shows this three-dimensional solution space spanning application areas vertically and team levels horizontally. This continuous documentation captures each player’s individuality and their development over years.
Figure 1: Solution Space Matrix.
The right mindset: Compass for digital transformation
To master complex challenges, a clear compass is needed. Holstein Kiel decided to rely on the Data Value Generator Flywheel. Figure 2 illustrates this cyclical approach that begins with recognized data potential and develops concrete use cases from it. The technical infrastructure is designed to enable rapid iterations, from initial prototypes to functional solutions, without wasting resources.
If a use case shows positive results, it can be quickly expanded thanks to the flexible architecture. With each successful project, the data landscape keeps growing along with measurable benefits for the core business. This creates trust, increases project visibility, and brings additional resources, both personnel and financial.
Figure 2: Data Value Generator Flywheel.
Practical example: Intelligent player analysis
To implement the described model, Holstein Kiel, together with AWS, defined five technical core principles.
These include:
- Scalability
- Standardization and automation
- Accessible and secure data access for all users
- Modular architecture
- Event-driven and logically linked functions
Scalability is the foundation for sustainable growth. AWS offers the advantage, providing infrastructure which grows with your organization. You only pay for what you use, in line with a pay-as-you-go model—no high upfront investments are needed. Standardized and automated data pipelines facilitate efficient data management, while modular architecture enables flexible expansions. This significantly reduces technical effort, especially important for a club that doesn’t primarily function as a technology company and wants to focus on its core business.
A scouting application would be the connection between physical attributes (such as acceleration speed and the number of accelerations) and a player’s goal threat (pXT). pXT is defined as the change in the probability of a specific player scoring as the ball moves closer to the opponent’s goal. Figure 3 shows an Amazon Quick Sight dashboard, which displays performance metrics including overall score, possession, and scoring percentages for individual players.
Figure 3: Amazon Quick Sight dashboard.
On this basis, holistic digital twins of both their own team and external teams from almost anywhere in the world can be differentiated. Holstein Kiel can quickly generate insights about which players could potentially be considered as transfer options to increase their own team performance.
Exploring future possibilities: OpenAI on Amazon Bedrock
Holstein Kiel is experimenting with AI technologies using OpenAI models on Amazon Bedrock combined with an intelligent knowledge database, Amazon Bedrock Knowledge Bases. This approach enables sporting decision-makers to communicate with player data through a natural language chatbot interface. Squad planners can ask complex questions about player performance and compare their own players’ characteristics with external players.
The solution automatically provides a selection of potentially interesting transfer options. These recommendations are based on a multidimensional player profile that includes sporting, economic, and personal factors. Technically, data from training devices automatically flows into a central data storage using Amazon Simple Storage Service (Amazon S3) and is prepared for AI analysis. The OpenAI models on Amazon Bedrock perform complex analyses that go beyond conventional queries. The AI-generated insights are then provided flexibly and intuitively on various devices.
Alexander Rudies, the Head of Match Analysis/Recruitment/Innovation stated his thoughts about the potential of AI-supported efficiency gains for future player analysis. “In a structure with limited personnel resources, it’s important to deploy the expertise of each person correctly at the decisive process steps in a targeted manner. AI will help us in the future primarily to take over a large part of our processes in operational work. This allows us to deploy human expertise more specifically on developing solutions and communicating information to our players.”
Conclusion: Insights and outlook
The initial data journey experiences of Holstein Kiel highlight important success factors. Flexibility is crucial to respond to the uncertainties of the football business. Current experiments with AI technologies, like OpenAI models on Amazon Bedrock, open completely new possibilities—for greater efficiency and for innovative business models.
Critical to success is the right mindset for cultural and organizational changes, as well as technical infrastructure that accelerates innovations, rather than slowing them down. The player-centric data architecture combined with experimental AI approaches positions Holstein Kiel as a pioneer for data-driven innovation in football.
This combination creates maximum resource efficiency and measurable business value. It leads to greater acceptance, higher visibility, and ultimately additional resources. However, the club’s main goal always remains: Holstein Kiel wants to win football matches and develop players, supported by the power of artificial intelligence and AWS services.
The experience of Holstein Kiel demonstrates how sports organizations can leverage AI and cloud infrastructure to gain competitive advantages through data. To implement a similar solution, start by identifying your data silos and defining player-centric data architecture. Learn more about Amazon Bedrock for building generative AI applications or explore AWS for Sports to see how other organizations are transforming team performance with data analytics.
To discuss how your organization can leverage AI and data analytics for a competitive advantage, contact an AWS Representative.