Ben Kus, Box | Supercloud 4
summary
Box explains how generative AI turns unstructured content into usable assets by pairing large foundation models (OpenAI, Google, Anthropic) with 90% of enterprise data is unstructured and using retrieval-augmented generation (RAG) with vector embeddings to retrieve relevant docs before prompting. Product work focuses on embedded features — document Q&A, automated tagging, structuring and permission-aware search — and argues ROI comes from increased productivity and new capabilities rather than headcount cuts, while watching token/inference cost vs TCO.
AI-written summary · about
chapters (14)
- 0:00Introduction to Supercloud 4
- 0:22Generative AI impact overview
- 1:06Unstructured data challenges
- 3:33Business value of AI adoption
- 4:56AI wrappers vs native systems
- 7:26AI systems in production
- 9:30Cost savings vs revenue growth
- 11:15AI as software enabler
- 12:18Data ownership and models
- 15:27Vector databases and embeddings
- 17:15AI cost and performance
- 19:23Box's AI product focus
- 20:48AI operating system vision
- 21:32Future of platform engineering
