The hardest part is just asking. Frame the course as a fix for a problem the team already has — shipping AI features without a way to measure quality. Copy the template below, fill in the brackets, and send it.
Subject: Professional development — AI Agent Evaluation course
Hi [Manager's name],
I'd like to enroll in AI Agent Evaluation, a 5-week course from softmax on measuring and improving the quality of AI agents before they reach production.
We're shipping more AI-powered features, and right now we evaluate them mostly by eye. This course covers exactly what we're missing: building eval datasets, error analysis, LLM-as-a-judge, multi-turn evaluation, and folding evals into CI/CD so regressions are caught before release, not by users.
What I'd take away and apply directly to our work:
• A repeatable process to measure agent quality instead of guessing
• Concrete ways to catch and fix the failure modes we ship today
• Practical patterns I can share with the rest of the team
It's taught live by a Head of Product who built AI products in production at Revolut. The cost is $1,000, it starts on August 10, 2026, and it runs alongside full-time work (about 4–6 hours a week).
Details: https://softmax.school
Would you approve expensing this against our L&D / training budget?
Thanks,
[Your name]