Background

Artificial Intelligence (AI) is rapidly becoming a core capability of enterprise systems such as ERP and cloud platforms, where it is used to automate processes, enhance decision-making, and enable new digital services. However, these advances come with growing computational and energy demands. The training and operation of AI models increasingly contribute to data center electricity consumption and associated greenhouse gas emissions, creating a tension between ambitions for digital innovation and corporate sustainability goals.

In response, the notion of “Frugal AI” has emerged, emphasizing AI solutions that deliver business value while minimizing energy use, computational complexity, and environmental impact. Rather than maximizing model size or accuracy at any cost, Frugal AI foregrounds efficiency, sufficiency, and conscious trade-offs as design principles. Enterprise software providers play a pivotal role in operationalizing this paradigm. As a leading provider of enterprise applications, SAP is integrating AI capabilities across its portfolio (e.g., S/4HANA, Business Technology Platform). This context offers a highly relevant setting to investigate how sustainability considerations are–or are not–embedded along the AI lifecycle, from development and architectural decisions to configuration and everyday use in customer organizations.

To understand how these lifecycle decisions are reflected in the enterprise system itself, the thesis will examine where they materialize across the user experience, process, foundation, and platform layers. The thesis will also link sustainability-relevant choices to concrete architectural elements while preserving the temporal and socio-technical logic through which AI sustainability emerges over time. Building on this, the thesis explores how frugal and sustainable AI can be operationalized in enterprise systems, with a particular focus on energy efficiency and environmental impact.

Research Goal

The thesis aims to examining energy efficiency and environmental impact across SAP’s User Experience, Process, Foundation, and Platform layers while considering how practices in development, configuration, and everyday use shape sustainability outcomes across these layers:

• User Experience Layer: How do user-facing design choices influence the footprint of AI-enabled enterprise applications and how do interaction patterns, defaults, transparency, and guidance influence AI usage intensity and user behavior? Which mechanisms can promote more frugal everyday use without undermining business value?
• Process Layer: How does embedding AI into business processes affect resource use, and how can it be configured sustainably? How are AI capabilities integrated into workflows, applications, and agent-based process automation, and which process design choices drive compute intensity? How do SAP and customer organizations configure and deploy AI-enabled processes to optimize resource use, and what socio-technical factors enable or hinder sustainable configuration?
• Foundation Layer: How do model-, data-, and knowledge-related decisions influence the system’s energy efficiency and environmental impact? How do sustainability considerations including energy-aware design and coding practices influence efficiency and environmental impact at the AI foundation level?
• Platform Layer: How do platform and shared services enable or constrain frugal AI?
• How do sustainability considerations including energy-aware design and coding practices influence efficiency and environmental impact at the AI foundation level? Which platform-level architectural and engineering decisions shape the energy footprint of AI capabilities (e.g., runtime, scaling, monitoring, observability, identity & access, integration)?

A cross-cutting objective is to understand how decisions at these four layers interact over time to shape overall sustainability outcomes.

Research Design

The thesis will conduct an interpretive, single-case study in collaboration with SAP and at least one customer organization adopting an AI-enabled SAP solution. The unit of analysis is an AI integration initiative spanning development at SAP and subsequent use at the customer.

Data collection can include:

• Semi-structured interviews with, e.g., SAP developers, product managers, sustainability experts, consultants, and customer-side IT staff.
• Internal and public documentation (architecture documents, guidelines, sustainability reports).
• Where feasible, system logs or usage statistics to complement qualitative insights.

Your Profile

• You are interested in the emerging field of AI and sustainability
• You are confident in conducting executive interviews and comfortable engaging with industry professionals
• You are highly motivated to work on real-world challenges in a self-organized and goal-oriented working mode, and you bring in your own ideas
• Very good English skills, as the thesis will be written in English

Details

• The thesis will be written in collaboration with SAP and will be remunerated
• Start: January 2026
• Language: English

We offer you an exciting research topic, close supervision, and the opportunity to develop both practical and theoretical skills. If you are interested, please send a current transcript of records, a short CV, and a brief motivation (2-3 sentences) to Linda Sagnier (linda.sagnier∂kit.edu).