Three announcements published the same day on the OpenAI blog outline a consistent strategy: converting consumer use cases into deep enterprise contracts. Uber is deploying OpenAI models across its real-time marketplace platform — on the driver side for earnings optimization, on the passenger side for voice booking. Singular Bank built Singularity, an internal assistant on ChatGPT + Codex that saves its bankers 60 to 90 minutes per day on meeting prep and portfolio analysis. This isn't a POC: it's embedded critical workflow. The B2B study published in parallel ("How frontier firms are pulling ahead") provides the narrative framework accompanying both deployments — Codex as an agentic automation tool to widen competitive gaps.
The underlying signal is less "AI helps companies" and more "OpenAI is consolidating a position as a cognitive infrastructure provider." Uber and Singular Bank aren't customers testing an LLM: they've built internal and external products on top of it. Migration costs become structural. Codex appears in both enterprise use cases most prominently featured today — that's not coincidental, it's the product OpenAI has been actively pushing in B2B since its April 2025 repositioning.
On the margins, the ChatGPT privacy publication (reduction of personal data in training, user control over conversation usage) lands in a context where enterprise deployments require precise contractual guarantees. It's a reassurance document as much as a technical explanation — useful to have on hand for legal and security teams evaluating these integrations.