TLDRocket
Sign in

HiBob turns 2,500 GPTs into product and team growth

OpenAI

HiBob rolled out 2,500 custom GPTs across its company using ChatGPT Enterprise. The twist: it's driving real revenue, not just internal shortcuts.

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

HiBob, the company behind the HR platform Bob, has quietly turned ChatGPT Enterprise into something closer to internal infrastructure than a chat window. Over the past year, employees across sales, support, product and HR itself have built roughly 2,500 custom GPTs, each tuned to a specific task rather than some generic all-purpose assistant.

That number matters because it signals scale, not novelty. This isn't a handful of engineers tinkering on a hackathon weekend. It's a company-wide habit where a support rep building a GPT to draft policy answers is treated the same as a product manager building one to summarize customer feedback. HiBob leadership pushed adoption hard, treating GPT-building almost like a literacy skill employees were expected to pick up.

The payoff shows up in two places. Internally, HR and go-to-market teams say routine work — drafting onboarding docs, triaging support tickets, prepping sales materials — moves noticeably faster now that a purpose-built GPT sits behind each workflow instead of a blank prompt box. Externally, some of that internal tooling has migrated straight into the Bob product itself, turning experiments into AI-powered features customers actually use, which HiBob credits with part of its recent revenue growth.

What's notable is the direction of the flow. Most companies talk about AI adoption as a top-down rollout: leadership picks a tool, mandates training, hopes for the best. HiBob's version looks more bottom-up, with thousands of small, employee-built tools feeding upward into both internal efficiency and the product roadmap. And that bottom-up mess, chaotic as 2,500 individual GPTs sounds, appears to be exactly the point.

My take — AI-written commentary, not fact-checked reporting

I've grown tired of enterprise AI case studies that amount to 'we bought a chatbot and morale improved.' This one at least shows a mechanism — thousands of narrow, employee-made tools compounding into product features and revenue — which is a far more believable story than another vague productivity claim. My bias: closed tools like ChatGPT Enterprise win these workplace deployments right now not because they're better models but because they ship with the boring plumbing — admin controls, security, integration — that open-weight alternatives still make companies build themselves.

Read more about this at: OpenAI

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.