SAP's Sean Kask: Software Firms Must Embrace AI or Risk Perishing
Tables and AI: The Core of Business Value
Tables hold most of a company’s value, and language models cannot read them, stated Sean Kask, chief AI strategy officer at SAP, referring to SAP’s significant investment in Prior Labs. Europe, according to Kask, should avoid aggressively pursuing large language models.
The Shift to AI-First
Kask emphasized the necessity for established software companies to integrate AI into their core products, not merely as an additional feature. He spoke at Wave by Vento in Turin, sharing insights on SAP's approach to AI.
Key Takeaways:
- Every software company must become an AI company to survive.
- 80% of business value is derived from 20% of data held in tables, making tabular models crucial.
- SAP's focus on tabular foundation models, like Prior Labs, allows for faster and more accurate predictions and classifications.
- SAP plans to maintain Prior Labs as a separate lab, encouraging open-source research and collaboration.
- Europe should invest in building AI-enabling assets rather than competing in the costly race for large language models.
SAP's Strategy
SAP, having completed the acquisition of Prior Labs in July, has committed over €1 billion over four years to develop the startup's tabular AI technology. Kask highlighted SAP's existing in-house tabular model, SAP-RPT-1, and its potential applications in diverse fields like cancer diagnosis and bank transactions.
The Future of AI in Business
Kask suggested that the future of AI lies in efficient, accurate models that can handle structured data like tables. SAP, through its partnerships and ecosystem, aims to create such models, ensuring they remain accessible and beneficial to all European companies.