OpenAI reports that Stampli, a B2B accounts-payable automation platform, cut the number of hours needed to launch new customer accounts by 68% after adopting ChatGPT Work, OpenAI's workplace product.
The case study frames the gain as coming from Stampli's implementation team using ChatGPT to handle tasks tied to standing up new customer accounts — the kind of configuration, documentation and repetitive setup work that typically consumes the early weeks of a B2B software rollout. OpenAI did not publish a detailed breakdown of which specific tasks were automated or how the 68% figure was measured; that methodology is unconfirmed beyond OpenAI's own summary.
What is confirmed is the direction: a company selling automation software to other businesses is using a general-purpose AI assistant to compress its own service-delivery timeline, not just its product's output. That distinction matters for operators evaluating AI tools, because onboarding and implementation are typically treated as too client-specific to automate, unlike ticket routing or lead qualification. Stampli's result suggests that assumption is being tested.
For consultancies and internal ops teams built around manual client setup, the immediate implication is not that ChatGPT alone replicates this result out of the box — Stampli's implementation presumably involved building internal workflows, templates or prompts around the tool rather than using it unconfigured. The case is best read as evidence that the effort-to-payoff ratio for automating onboarding has shifted, not as a plug-and-play recipe.
Companies with implementation or customer-success teams of any size should treat this as a prompt to time-audit their own launch process: which hours are spent on decisions unique to each client, and which are spent re-writing the same setup steps, config summaries, or client-facing documentation with minor variations. The latter is exactly the category OpenAI's case study targets, and it is the category most B2B operators have not yet touched.