AI training for software teams
A live AI agent on real tickets, with a course for each role: developers, QA, product and engineering managers
For your team
What they learn. What it's worth.
Who
What they learn
- PromptingA precise brief, not “fix the bug”.
- Cost managementLower AI cost per ticket: search, don’t paste.
- PlanningAgree on a plan before the agent writes.
- Model choiceThe right model for each task.
- Data privacyKeep personal data out of prompts.
- HallucinationsCheck AI’s facts against the source.
Each role’s course has its own mix. See more →
What it's worthYearly · USD
AI cost per solved ticket, sample team after Course 1 Sample team. AI cost per solved ticket (tokens × model price), same tickets, before → after the lessons.
Savings per year Estimate: employees × AI cost × 12 months × 30%, if your AI cost per solved ticket falls as it did in the sample team.
AI cost today, per year$360,000
How we teach
Try. Get feedback. Retry.
1
Try2
Get feedback3
Retry4
See your score improve@src/cart.ts … expected €45Demo run: one person, one task, first try and retry. Same grader both times.
Developer
QA
Product manager
Engineering manager
Courses per role
Pick your role. Camp 1 is free.
What does it cost?
A one-off price per person per course, excl. VAT: €79 each for the first 15 people, €69 for the next 35, €59 for the next 150 and €49 above 200. Camp 1 of every course is free.
How much time does it take per person?
About 10 to 15 minutes per camp, so 1 to 2 hours for a whole course of 6 to 8 camps (our estimate until enough people have finished to measure it). Each person goes at their own pace.
Which AI tools does it use?
Learners direct a live AI agent running Anthropic’s Claude models (Haiku, Sonnet and Opus) in a Claude Code-style session. The skills, such as precise prompts, planning first and picking the right model, apply to any AI tool your team uses.
Do people need to install anything or bring an AI licence?
No. Everything runs in the browser, including the AI agent and its costs. Nothing touches your code, accounts or tools.
What happens to our data?
Our servers, database and backups are in the EU (Germany). AI requests go to Anthropic to run the agent; we keep only token counts, never prompt text. No third-party trackers. We sign a DPA (data processing agreement) on request.
What does the manager see?
A team report with each person’s score per camp, first try and best, exportable as CSV. Everyone who finishes a course gets a certificate with a link anyone can check.
Who is it for?
Developers, QA, product managers and engineering managers. Each role has its own course set in its own work: developers and QA in a code repo, managers and PMs in a documents workspace.
How is it different from a video course?
Nobody watches slides. People solve real tasks with a live agent, and a grader scores how they worked, the way a code review would. The first try and the retry show what changed.