Situation
inMusic supports a family of world-renowned music technology brands across every time zone. The support team faced high volumes of repetitive questions about shipping, order status and product specifications, long queues at peak hours and the need to answer in many languages, which left little time for the technical issues that really need a specialist.
The team spent over a year testing AI tools, both native to Freshdesk and external. None met their bar for technical depth or for the empathetic, human-like experience they wanted.
Implementation
Riff was trained on more than 2 700 sources: product manuals, knowledge base articles and product pages covering specifications, troubleshooting, firmware and documentation for all brands. When a customer asks, Riff finds the relevant article and walks the customer to the answer instead of improvising.
Freshdesk is the hub of inMusic's support, so the hand-off had to be seamless:
- Riff answers and diagnoses, collecting the relevant details and running through the basic troubleshooting steps.
- If the issue cannot be solved in chat, Riff escalates to a human agent.
- A Freshdesk ticket is created automatically with a summary of the whole conversation, so the agent starts with full context.
inMusic set up, beta-tested and went live within two weeks. Improvements were made on the live instance, without a sandbox, with quick answers from the ChatLab team.
Result
According to inMusic's internal support metrics:
- 35%+ of support conversations resolved instantly in chat,
- 46% shorter average first response time, with agents now answering within 8 business hours,
- CSAT peaked above 90%, an all-time high,
- fewer than 1 in 5 conversations end in a Freshdesk ticket,
- 14 000+ conversations a month.
Customers in post-service surveys describe the AI support as precise and professional, and the hand-over to a human as smooth. Agents jump straight into a personalised answer without going over old ground.
