Rosa's Wednesday, two ways

Rosa, a shift meteorologist at a regional forecast office.

Rosa starts her Wednesday at 5 a.m. with the overnight model runs and a line of storms moving toward the river. She opens ChatGPT and uses the daily forecast drafting prompt, pasting in the model trends and her own notes. The first draft of the forecast discussion comes back in a minute, and she edits the timing and the local details.

By mid-morning, the storms look stronger. She uses the severe weather messaging prompt to draft warning text and safety instructions, then checks every word against the radar. She sends a short briefing to the county emergency manager using the emergency services coordination prompt, with an impact timeline for the afternoon.

After the storms pass, she uses the verification and post-event report prompt to compare what happened with what was forecast. She notes the hail reports from spotters and the wind gust at the airport. The report is ready before her shift ends.

Rosa finishes on time for the first time in weeks. She uses the extra hour to walk by the river and to sit with a junior meteorologist who is learning the warning process. The prompts did not make the decisions. They made room for her to make them.