Organizations are rapidly embedding AI tools into their workflows, and managing AI providers has become a distinct and complex challenge. Familiar sourcing tensions around cost, data sovereignty, vendor independence, and regulatory compliance apply here as in traditional IT procurement, but AI introduces additional complexity that existing frameworks do not adequately address. Provider outputs are probabilistic and non-deterministic, models drift or are deprecated without notice, and hallucination risk varies across providers and use cases. The provider landscape itself is broad and evolving quickly, spanning dedicated model providers, hyperscalers, and AI-enhanced additions to existing software ecosystems, with new access modalities emerging continuously. Organizations navigating this landscape lack a structured basis for comparing and selecting AI providers in a systematic way.

The thesis is supposed to answer the research question: "What dimensions should a structured AI provider comparison framework incorporate to support organizations in strategic sourcing decisions?"

If you are interested in this thesis, feel free to reach out to Dominic Sieber (dominic.sieber@unisg.ch).