Building a Data Layer for AI | Juan Bruce

Juan Bruce, co-founder of Atomic Strata, discusses how enterprise AI is shifting from model choice to governed, auditable data and context. He says EU regulation will push enterprises to log the data feeding AI outputs, while closed AI stacks from OpenAI, Anthropic, and Google create lock-in. He also argues stablecoins will likely become the payment rail for AI agents. - Juan Bruce leads Atomic Strata, which builds an enterprise-grade AI context/data layer. - He argues AI output quality is increasingly driven by data and context rather than the model alone. - EU AI regulation is expected to require audit logs of the data and context used in AI outputs. - IBM and large system integrators are seen gravitating toward open-source AI, while enterprises face lock-in with closed AI stacks. - Stablecoins are framed as the likely payment rail for AI agents, with smart contracts enabling gated execution. - Sixt is cited as an enterprise allowing employees to build AI apps within governance guardrails. - Juan uses multiple AI products and a shared context layer to avoid re-explaining work and prevent developer conflicts.