
Not Diamond, a leading multi-model AI infrastructure provider, unveiled its Prompt Adaptation system at SAP Sapphire 2025 on May 20, 2025, in San Francisco. This agentic system automates prompt optimization across multiple large language models (LLMs), enhancing enterprise AI efficiency and scalability. The launch, accompanied by new funding from IBM, SAP.iO Fund, and others, positions Not Diamond as a key player in the multi-model AI landscape.
Not Diamond launches Prompt Adaptation at SAP Sapphire 2025.
Automates prompt optimization across multiple LLMs for enterprise AI.
Integrated into SAP’s generative AI hub for enhanced model agility.
Secured funding from IBM, SAP.iO Fund, Myriad Venture Partners, and others.
Reduces manual prompt engineering by thousands of hours.
Enables real-time model selection for cost and performance efficiency.
Prompt Adaptation is an agentic system that programmatically rewrites and optimizes prompts across various AI models, eliminating manual engineering efforts. Integrated into SAP’s generative AI hub platform (SAP AI Core and SAP AI Launchpad), it was previewed by SAP CTO Philipp Herzig, showcasing its ability to drive accuracy and streamline engineering workflows. “Prompt optimization is key to make Enterprise AI more flexible, cheaper, more accurate, and less brittle,” Herzig stated. The system reduces model onboarding time from months to minutes and cuts manual prompt engineering by thousands of hours.
Not Diamond’s collaboration with SAP enables enterprises to adapt prompts dynamically to the best available model, ensuring optimal performance across use cases, regions, and regulatory requirements. “With Not Diamond, we can adapt to AI innovations and new models much faster and benefit from performance improvements in AI development,” said Herzig. This integration supports SAP’s customers in navigating the rapidly evolving AI model landscape, enhancing flexibility and reducing costs.
The launch coincides with a new funding round led by IBM, SAP.iO Fund, Myriad Venture Partners, Deepwater, DNX, Ambush Capital, and Defy. This investment underscores the growing importance of multi-model orchestration and prompt adaptation. “Generative AI requires a fit-for-purpose model strategy,” said Emily Fontaine, VP at IBM. “Not Diamond combines deep research with real-world impact. Their routing and prompt adaptation systems help scale AI.” The funding will support Not Diamond’s mission to build infrastructure for a modular, dynamic AI future.
Prompt Adaptation leverages intelligent workflows to test and refine prompts, offering enterprises:
Automated Prompt Optimization: Adapts prompts across LLMs for up to 25% improved accuracy.
Reduced Engineering Overhead: Saves thousands of hours by automating prompt engineering.
Dynamic Model Routing: Selects the best model for each task in 60ms, balancing cost and performance.
Security and Compliance: SOC-2 compliant with zero data retention and VPC deployment options.Posts on X highlight early customer success, with enterprises reporting faster development cycles and improved output quality.
Not Diamond’s Prompt Adaptation addresses the scalability challenges of manual prompt engineering, enabling enterprises to adopt new models rapidly. “We’re entering a new era of multi-model AI infrastructure,” said CEO Tomás Hernando Kofman. “Prompt Adaptation makes prompts programmable, improving accuracy and driving efficiencies.” With its integration into SAP’s ecosystem and robust investor backing, Not Diamond is poised to redefine how enterprises deploy AI at scale.
Not Diamond, is a multi-model AI infrastructure platform that helps enterprise AI/ML teams seamlessly assess, onboard, and optimize new AI models through cross-model prompt adaptation and intelligent model routing. Instead of waterfalling development inefficiently through one model at a time, we help companies save tens of thousands of engineering hours and improve output accuracy by evaluating and optimizing all candidate models and parameter configurations in parallel through our data-driven platform.