What is changing at work
New OpenAI research says AI is expanding the range of tasks people handle at work. In practice, people using ChatGPT are taking on activities that sit across traditional role boundaries, which changes how work gets divided and coordinated inside organisations.
Why this matters for business operations
When AI expands what people can do, teams should expect knock on effects in everyday operations. Work that used to be queued to specialists can move earlier in the process, and routine parts of knowledge work can be reassigned to the person closest to the workflow. That can improve throughput, but it also increases the need for consistent review, documentation, and access to the right information.
What teams should do next
- Map one or two real customer facing or internal workflows where role boundaries are currently a bottleneck, then identify which steps could be supported by AI without skipping required approval.
- Define what AI can draft and what humans must verify, especially for anything that affects customer communications or operational decisions.
- Pilot with a small group and track outcome based metrics tied to the workflow, such as time to first draft, reduction in rework, or speed to resolution.
- Update internal guidance so people know how to use AI outputs, how to cite or reference internal sources, and how to escalate issues when the model is uncertain.
Bottom line: treat this as a workflow redesign problem, not a prompt learning problem. The value comes from changing who does which steps, then controlling quality and measuring the result.