How hepster reduces the burden on Customer Support with octonomy

hepster is an InsurTech company specializing in embedded insurance: insurance solutions that integrate seamlessly into digital customer journeys, including mobility, electronics, sports, pets, travel, furniture, and luxury goods.

For partners, hepster relies on API-based integration, enabling insurance to become part of existing processes quickly and efficiently.

As hepster’s product and partner business grows, so does the variety of incoming requests, and with it, the pressure on customer support, including the need for fast response times, accurate answers, and clean documentation.

hepster extended its customer service capabilities with octonomy’s AI expert solution, which understands requests, draws on information from systems and knowledge sources, and completes processes end to end.

50% reduction in customer support workload – with no loss in quality.

Operating an embedded insurance business means navigating a wide range of contexts, rules, and edge cases – from product details and contract or pricing logic to documents and verification requirements. At the same time, customers expect fast, precise responses across digital channels.

Traditional chatbots are often limited. They have no system access, no reliable auditability, and no real process handling. That gap was the starting point for hepster. Automate, but in a way that does not compromise quality, compliance, or the customer experience.

hepster focused on three core objectives:

1. Day-to-day relief – less manual routine work and more capacity for complex cases.

2. Consistent answer quality – reliable, accurate responses instead of the unreliable answers typical chatbots produce

3. Scalability – new topics, products, and partners should not drive linear growth in support headcount.

octonomy operates as a multi-agent system, not a single all-purpose tool.  It’s a coordinated team of specialized agents (e.g., for contract inquiries, complaints, payments, and documents), complemented by a Supervisor and Ops Agent responsible for planning, routing, and documentation.

• End-to-end resolution, not just FAQs: Understand the request, locate and verify relevant information (including documents), and close out the case cleanly, including providing a response back to the customer.

• Integration with existing systems: Connected to CRM, ticketing systems (e.g., Zendesk), policy/billing systems (planned), and more via APIs and connectors.

• Auditability & quality: Every step and decision can be logged with responses strictly based on verified knowledge sources.

• Enterprise-ready: Multilingual, scalable, and GDPR-compliant data processing in accordance with European standards.

hepster took a pragmatic approach: defined 1–2 clearly scoped use cases, established KPIs (e.g., workload reduction, response time, resolution rate), launched a pilot, and then expanded iteratively.

• Channels: Email / Chat / Contact Form

• Systems: Ticketing system, CMS, knowledge base / policy documents

• Knowledge sources: General Terms and Conditions (T&Cs), product materials, internal guidelines, process documents

• Rollout strategy: automate chat interactions on the website and routine contact form requests; hand off complex cases to human agents with full context (combining human and AI intelligently)

• Channels: Email / Chat / Contact Form / + Customer Account

• Systems: Ticketing system, CMS, knowledge base / policy documents / + Policy System

• Knowledge sources: T&Cs, product materials, internal guidelines, process documents

• Rollout logic: Once quality is approved → direct replies to ticket requests; execute standard requests, such as address or name changes, via the Admin Ops Agent

With a scalable automation layer in place, additional processes can be added step by step, including status inquiries, master data and contract changes, and general information delivery (with varying levels of complexity depending on the product line and sales channel).

Particularly during phases of growth and new partnerships, such as portfolio expansion or internationalization, the goal remains clear: scale without linear headcount growth.