Stampli cuts launch hours 68% using OpenAI's Codex
Stampli, a procure-to-pay software platform, is the subject of a new OpenAI customer case study describing how its product marketing team used Codex and ChatGPT Work to launch a new product, Deep Finance, on a fixed deadline while design resources and outside contractors were already committed elsewhere. Deep Finance turns data from Stampli's procure-to-pay platform into spend intelligence for CFOs, VPs and other business leaders.
The team connected product context, meeting notes, decisions and messaging guidelines into a shared system built with Codex, then used it to produce a seven-part blog series, launch emails, a webinar and its supporting deck, social and paid creative, a PR Newswire release, the Deep Finance web page, and sales enablement materials. Codex also handled roughly 90% of the polished hero animation for the launch, covering exploration, iteration and packaging, before a contractor finished the opening scene and final format. Stampli estimates the full go-to-market and content workflow would have taken about 243 modeled active role-hours without Codex; with it, the work took approximately 77 hours, a savings of roughly 166 hours and a 3.16x speedup, which the case study's headline rounds to a 68% cut in launch hours. Human review and final approval stayed in place on everything customer-facing. Deep Finance itself went from an initial prototype demo to public launch and first shipped product in about six weeks, which Melad Zahedi, Stampli's Director of Product Marketing, says previously would have taken months or quarters.
The same Codex-based system now runs day to day: it pulls information from product systems and meeting notes to keep help center articles, presentations and one-pagers current, work that used to mean interviewing product managers, reading Jira tickets and reviewing GitHub by hand. Zahedi says the automation has multiplied his small team's output by 10x, taking it from a couple of pieces of content a week to hundreds. He also uses ChatGPT Work as a 'second brain' to prepare for a schedule of back-to-back meetings, and in one executive meeting an employee used Codex to pull and analyze HubSpot metrics live, producing in about 20 seconds of keystrokes an answer Zahedi says would otherwise have taken Stampli's FP&A team half a day.
Key facts
- Stampli estimates its Deep Finance launch took about 77 modeled active role-hours with Codex versus about 243 without it, a 3.16x speedup the case study's headline rounds to a 68% cut.
- Codex produced a seven-part blog series, launch emails, a webinar deck, paid creative, a PR Newswire release and the product web page, and handled about 90% of the launch's hero animation.
- Deep Finance went from prototype demo to public launch in about six weeks, which Director of Product Marketing Melad Zahedi says previously would have taken months or quarters.
- Zahedi says GPT-powered automations have multiplied his product marketing team's weekly content output by 10x, from a couple of pieces to hundreds.
- In one executive meeting, an employee used Codex to pull and analyze HubSpot metrics in about 20 seconds of keystrokes, work Zahedi says would have taken Stampli's FP&A team half a day.
Why it matters
This is an OpenAI-published case study, not independent reporting, but it gives a concrete, numbered picture of how one company folds Codex and ChatGPT Work into an entire product marketing function rather than a single task: content production, meeting preparation, ad hoc data analysis and day-to-day knowledge management are all described as running through the same system.
Who it affects
Directly, Stampli's product marketing team and its Director of Product Marketing, Melad Zahedi, who is quoted throughout. More broadly, it is aimed at marketing and go-to-market teams at other companies, and at OpenAI's enterprise customers weighing ChatGPT Work and Codex for similar workflows.
How to use it
The source describes a workflow rather than a product feature: Codex connects product context, meeting notes, decisions and messaging guidelines into one system that marketing then draws on for blog posts, emails, decks, creative, PR releases, web pages and sales materials, plus ad hoc data pulls from systems like HubSpot. The case study gives no pricing or licensing detail for ChatGPT Work or Codex.
How solid is it
The figures come entirely from Stampli's own self-reported estimates as relayed in OpenAI's customer story: the 243-hours-versus-77-hours comparison is explicitly described as 'modeled active role-hours,' not measured time tracking, and the 68% headline figure is a rounding of the 3.16x multiple. As a vendor case study built to showcase OpenAI's own products, it carries no independent verification of these numbers.
Risks and caveats
The source gives no calendar date for the launch, no figures for Stampli's headcount or revenue, and no quantified baseline for content output before adopting these tools beyond 'just a couple' of pieces a week. The identity of the contractor who finished the hero animation is not given, and one quote about Codex making teams '10x faster from requirement to deployable solution' appears in the source without an attributed speaker.
“it's multiplied the output of a small team by 10x, putting out hundreds of pieces of content on a weekly basis, where it was limited to just a couple before.”
— Melad Zahedi, Director of Product Marketing at Stampli