The Claude Program's First Night, and What People Built by Morning

The Claude AI Program ran its first session on the evening of Wednesday, 1 July 2026.

The ground covered was deliberately unglamorous: a framework for prompting, how to fact-check what the model gives you back, and Claude’s artifacts. Nothing there is a trick. It is the groundwork you need before anything else you do with the tool is trustworthy.

What surprised us was the speed. By the next day several participants had already produced real reports and dashboards for their own professions. Not exercises from the slides. Their actual work, the thing sitting on their desk that week.

We think the fact-checking piece is why. When people are taught to verify output as a habit from session one, they trust themselves to use the tool on something that matters instead of keeping it at arm’s length for low-stakes drafts. Confidence comes from knowing how to check, not from being told the tool is good.

Putting prompting and verification in the same first hour is deliberate, and it is the part we would keep if we had to cut everything else.

New cohorts start regularly, they are built for beginners, and no coding is needed. Dates and details are on the Claude AI Program page.