Customer story
How Jurafuchs automates growth marketing and customer success with AI agent teams
300+ customer conversations a month, answered in German, with every reply reviewed by a human. The app that German law students use to master the law runs its customer support, growth, and operations with a team of AI employees on AgentTeams.
300+
support conversations handled per month, in German
2,000+
messages handled per month across all channels
3
AI teammates covering support, growth, and operations
12
tools connected, from the support desk to the SQL database
About Jurafuchs
Jurafuchs is Germany's leading learning app for law. Students and professionals use it to work through the law case by case, and the team behind it keeps a promise that leaves no room for sloppiness: the content has to be correct, current, and precise. A small team, a growing user base, and channels that never sleep: the support inbox, the growth reporting, the day-to-day operations behind a product students rely on every morning.
The challenge
Support requests arrive in German, they need real product and domain context, and the quality bar is absolute: a wrong answer about the law is not an option. That ruled out the usual chatbot. Growth work had the same shape: real data lives in the team's SQL database and analytics tools, and pulling reports together eats hours. What the team wanted was closer to new colleagues: teammates who do the work, follow house rules, and show their drafts before anything reaches a customer.
The AI team
Jurafuchs runs three AI teammates on AgentTeams, each with a name, a role, and their own responsibilities:
Susanne Porth · Customer Support
Works the support inbox end to end: reads, triages, tags, and drafts every reply in German.
Gabe Rowth · Growth Marketing
Works across analytics, campaigns, and reporting, connected to the team's data and marketing tools.
Sofia Bennett · Project Manager
Coordinates the team, curates company knowledge, files improvement proposals, and reports back daily.
Together they handle more than 2,000 messages a month across support, Slack, and internal channels, on 12 connected tools. Susanne drafts customer replies, Gabe pulls growth numbers straight from the database and turns them into reports, and Sofia keeps knowledge current and proposes what to improve next.
How support works now
This is the part every buyer asks about, so here it is step by step:
- A customer writes in. Susanne reads the request in German and identifies what it is actually about.
- She triages and tags it against the team's nine-category schema, and only on explicit evidence in the ticket. No guessing.
- She drafts the reply. Nothing goes out automatically. Every draft lands with the team for review, and approval is one click.
- Spam, cold outreach, and automated notices get filed away correctly without ever interrupting a human.
More than 300 support conversations a month run through this flow. The team reviews drafts instead of writing replies from scratch, and the customer always hears a human-approved answer.
"Our quality bar leaves no room for error: wrong answers about the law are simply not an option, which is why classic chatbots never made it past a demo here. What convinced us about AgentTeams is the division of labor. Our AI teammate does the reading, the sorting, and the drafting, and a human gives the final approval. We kept our standards and lost the busywork."
The stack
Help Scout for the support desk, Slack for review and collaboration, the team's own SQL database and analytics for growth work, and the Jurafuchs knowledge base so every teammate answers with real company context.
Your next teammate starts with a conversation
Describe a process from your company and get a complete role proposal: responsibilities, tools, and starting knowledge included. No account needed.
Design your first role