Surging developer productivity with custom GPTs
OpenAI
Gaming company Paf rolled out ChatGPT Enterprise to its whole workforce, and engineers now lean on custom GPTs daily for coding grunt work. The bigger tell: 70% of all staff use it, not just devs — this is company-wide habit, not a tech-team experiment.
Based on reporting by OpenAI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Paf, the gaming operator behind the grit:lab coding academy, has gone all in on ChatGPT Enterprise, and the numbers OpenAI is touting say something about where this technology actually lands once the novelty wears off. Seventy percent of Paf's employees use it regularly, and that's not confined to engineering. Finance, HR, marketing, customer support — the tool has quietly become part of how a mid-sized company runs, not just how its developers ship code.
The engineering angle is still the headline, though. Paf's developers use custom GPTs day in and day out to chew through routine tasks, the kind of repetitive scaffolding work that eats hours without teaching anyone anything new. That's the pitch OpenAI wants enterprises to hear: less time on boilerplate, more time on the actual hard problems.
What's more interesting is the grit:lab piece. Paf runs its own coding academy at gritlab.ax, and it has folded ChatGPT Enterprise straight into the curriculum. Students aren't learning to code and then bolting AI on top later as a workplace habit — they're being trained from day one to think in terms of systems architecture with an AI collaborator already in the loop. That's a meaningfully different starting point than how most of today's senior engineers learned the craft, and it's a preview of what
My take — AI-written commentary, not fact-checked reporting
I'd bet grit:lab's approach ages better than most corporate AI rollouts, precisely because it's teaching juniors to architect systems with AI as a given, not as a bolted-on shortcut — that's the habit that compounds. The 70% adoption number is the real story here, though; it's a reminder that the boring back-office use cases, not flashy coding demos, are where enterprise AI actually earns its budget line.
Read more about this at: OpenAI