Our client, a leading German telecommunications company, wanted to translate AI ideas into managed products in a standardized way and identify opportunities for cost savings and additional revenue. Together, we developed an Alfabet-based approach that links requirements, use cases, and AI products with portfolio governance, technical architecture, and business value.
01 Background
Why AI Products Need More Than Just a List of Ideas
AI initiatives often emerge within different teams, with varying levels of detail and through different implementation paths. Without a common process, teams may pursue similar ideas independently of one another and overlook data or models that could be reused. As a result, decision-makers lack a complete view of costs, benefits, risks, and technical dependencies.
Our client therefore needed more than just a list of AI use cases. The goal was to establish a clear path from the initial idea to a product that could be implemented, operated, further developed, and eventually replaced. Throughout this process, responsibilities, governance requirements, and the connection between business needs and the technical product needed to remain clearly visible.
02 Procedure
A structured path from concept to portfolio
Identifying Potential
New AI ideas are submitted to Alfabet through a standardized demand process. Each demand documents the responsible parties, the requesting IT unit, stakeholders, operational responsibility, and the expected business value. The status indicates whether the idea is open, under evaluation, under review, approved, or closed.
Evaluation and prioritization take place in Demand. The result is clearly documented: The idea is either rejected or moved forward as an approved AI use case. This creates a transparent transition from exploration to implementation without the need for a separate shadow list.
Designing the AI Product
The operating model distinguishes between three interconnected elements. The demand accompanies an idea from discovery through to decision-making, while the AI use case describes how AI supports an approved business need. The AI product is the solution that is ultimately implemented and operated.
This distinction is important because an AI product can support multiple use cases. The product has its own architecture, responsibilities, and lifecycle, while each use case retains its specific business purpose, benefits, and governance context. Products and use cases are assigned separate identifiers but remain connected through relationships.
Alfabet links each product to its use cases, data products, technology platform, hosting components, AI models, and supporting technologies. Teams can identify dependencies, opportunities for reuse, and the parties responsible for operations.
Manage the Entire Life Cycle
The operating model for AI products includes Discovery, Delivery, Maintenance, and Decommissioning. Discovery begins with an initial assessment and leads to a detailed review of business and technical feasibility, while Delivery encompasses development and rollout. During operation, Monitoring, troubleshooting, and operational support are part of Maintenance.
Governance is embedded in every phase and begins with an early assessment that filters ideas before significant effort is expended. Product councils document decisions regarding further review and implementation, while portfolio, project, data protection, cybersecurity, and AI risk processes map out the necessary approvals.
The Playbook also provides a structured framework for prioritization. It takes into account strategic relevance and expected benefits, as well as the reliability of assumptions, reusability, implementation costs, and the workload for IT and AI teams. This allows limited resources to be focused on initiatives with demonstrable benefits and a realistic implementation path.
Manage products in the portfolio
Alfabet reports provide an overview of the AI portfolio and highlight relationships between use cases, products, data, and technologies. Color coding highlights relevant attributes, while Kanban views illustrate progress within the process. Business value is presented alongside technical and governance information.
Decision-makers can see which opportunities are being evaluated, which products are in development or in operation, and which use cases utilize the same building blocks. This allows teams to compare similar initiatives before separate products are developed. At the same time, shared data, platforms, and models become visible, so that reuse becomes a deliberate portfolio decision.
The product combines business value with implementation. It turns an approved AI use case into a solution that can be managed, operated, and further developed.
03 Result
A Common Foundation for the AI Portfolio
The implemented Alfabet model provides our client with a structured foundation for managing AI demands, products, and the entire portfolio. Instead of isolated entries, ideas, use cases, and AI products remain linked to the relevant data, models, technologies, responsible parties, risks, and their business value.
The Operating Model Playbook defines the overarching process that forms this foundation. Even as individual elements continue to evolve, the direction is clear: a transparent path from idea to a managed AI product, where decisions and dependencies remain visible throughout the entire lifecycle.
AI portfolio management is one of the use cases we implement in Bizzdesign Alfabet. Discover our Alfabet expertise or talk to us about your operating model.