In a case study published by OpenAI on August 20, 2026, procure-to-pay software company Stampli described using ChatGPT Work and Codex to launch its Deep Finance product. According to the account, the company moved from prototype demo to public go-to-market launch in about six weeks, a process it says previously would have taken months or quarters. Stampli's marketing team used Codex to turn product decisions and meeting notes into review-ready assets, including a seven-part blog series, launch emails, a webinar, social and paid creative, a PR Newswire release, a web page, and sales enablement materials. Codex also handled roughly 90% of a hero animation before a contractor finished the opening scene. Stampli estimates the launch content workflow would have taken about 243 modeled active role-hours without Codex and took approximately 77 with it — a saving of roughly 166 hours, or 3.16x faster production. Melad Zahedi, Director of Product Marketing, is quoted saying the tools 'multiplied the output of a small team by 10x.' The company says all customer-facing work kept full human review and final approval. The figures are Stampli's own estimates, reported by OpenAI as the primary source. OpenAI is the only source cited.
- Stampli estimates content production dropped from ~243 hours to ~77 (3.16x faster) using Codex
- Deep Finance moved from prototype to public launch in about six weeks, per the company
- All customer-facing assets kept human review and final approval, according to Stampli
- Figures are the vendor's own estimates, published by OpenAI on August 20, 2026
What it means for you
A software company says it used OpenAI's tools to do the busywork of a product launch — blog posts, emails, decks, social copy — far faster than before, while people still checked and approved everything. Treat the specific numbers with caution: they're the company's own estimates, published by OpenAI to promote its products, not an independent audit. The useful takeaway is the pattern, not the multiplier: connecting your scattered notes and product info into one place an AI can draw from can genuinely speed up repetitive content work.
Try this
Pick one recurring content chore — say turning meeting notes into a customer-facing update — and try having ChatGPT draft it from your notes this week, with you editing and approving the final version before anything ships.
Who should care
Small marketing and product teams that produce a lot of written and design assets on deadlines and feel understaffed for it.
Skip this if
You don't produce marketing or launch content regularly, or you're looking for hard proof of ROI — this is a vendor case study, not evidence for your specific situation.
Sources: OpenAI — read the original