Inside GPT-5 for Work: How Businesses Use GPT-5

OpenAI's latest report pulls back the curtain on something most AI coverage skips entirely: what people actually do with these tools once the novelty wears off. Instead of another benchmark chart, this is a look at real usage patterns across companies, broken down by department and task type. And the picture that emerges is less about AI writing code that ships itself and more about AI quietly absorbing the unglamorous parts of everyone's job.

The report frames adoption as uneven but accelerating. Some teams, particularly in engineering and customer support, treat GPT-5 as a daily tool baked into their workflow. Others are still experimenting, using it for one-off tasks like summarizing a meeting or rewriting a pitch deck. OpenAI's framing suggests this gap is closing fast, with usage spreading department by department rather than arriving all at once across an entire company.

What's notable is the shift in what people ask for. Early chatbot use skewed toward simple Q&A—define this, explain that. The tasks OpenAI highlights now look more like actual work products: drafting reports, analyzing data sets, generating first-pass code, structuring project plans. That's a meaningful jump from novelty to utility, and it tracks with what a lot of workers have been saying anecdotally for months—that the tools finally feel useful enough to build habits around, not just poke at out of curiosity.

The departmental breakdown is where the report gets interesting. Support and success teams lean on GPT-5 for handling ticket volume and drafting responses. Marketing and sales use it for content generation and prospecting research. Engineering teams use it for debugging and documentation, tasks that are tedious but not exactly glamorous. None of this is revolutionary on its own, but taken together it paints a picture of AI as infrastructure rather than spectacle—less a chatbot you talk to, more a layer sitting underneath how work gets done.

OpenAI's closing argument is that this is just the early innings, with usage set to deepen as trust builds and workflows get redesigned around the tool rather than bolted onto existing ones. That's a predictable thing for OpenAI to say, but the underlying data—real companies, real departments, real tasks—makes it harder to wave off as marketing spin than the usual splashy launch post.