Valeo Generates 35% Code With Gemini

Google Cloud just rewrote the playbook on deploying AI in business, compressing what used to be years of rollout into mere months—and signing rapid-fire expansions with Accenture, SAP, Oracle, and more. The big promise: moving from AI tools to true “agents” that execute real business workflows, not just fancy demos. SAP and Accenture are going all-in, with early access to DeepMind models and plug-and-play AI agents designed to drive billions in impact, while marketers can soon run complex, multi-step campaigns with a single prompt thanks to deep data integration between SAP and BigQuery. But the real test isn’t flashy features; it’s whether these agents actually improve outcomes instead of just automating bad decisions.

Here’s the catch: it all comes down to data orchestration and governance. If enterprise data stays fragmented and messy, even the best AI just moves faster in the wrong direction. Google’s answer is a technical deep-dive—bi-directional metadata flows with Collibra to ensure business context and policies travel with the data, and a marketplace of “sovereign” agents to address Europe’s strict regulations. Meanwhile, Valeo is showing industrial adoption at scale, rolling out Gemini agents to 100,000 employees and already generating over a third of its codebase with AI.

Featuring insights from Julie Sweet, Thomas Kurian, and reporting from TechCrunch, SAP News Center, and Vodafone.

Powered by Apisod.com