Customer story
CarOnSale puts customer support knowledge on autopilot with Qontext
CarOnSale is a B2B marketplace for the European used-car trade. Its customer-facing AI assistant helps buyers and sellers navigate auctions, payments, transport, complaints and other operational questions. Qontext gives the assistant fast access to relevant process knowledge and keeps it current as support documentation changes.
<1s
Hybrid retrieval time
0
Retrieval failures recorded
Why Qontext?
CarOnSale initially included support documentation directly in its assistant's prompt. As the knowledge base grew, prompt limits constrained the setup, and content updates required code changes and deployment. The team chose Qontext for fast access to relevant, continuously updated knowledge, without building and maintaining its own retrieval infrastructure. The setup also met CarOnSale's data processing, access control, security and EU data residency requirements for production.
Fast retrieval for customer conversations
Live customer conversations need fast access to context. Qontext keeps support knowledge normalized and indexed and makes it available through purpose-built retrieval endpoints. Retrieval takes under one second, giving CarOnSale's assistant the speed it needs to respond in real time.
Knowledge structured for agents
Qontext structures support knowledge for agent use. Its hybrid search retrieves the most relevant FAQs, approved wording and operational processes, even when a customer's question uses different terminology.
A self-updating flow
Confluence remains connected as a data source, while Sync and continuous updates keep the context repository current. When a support rule or process changes, the new guidance becomes available to the assistant without an application code change or deployment.
Qontext at CarOnSale
Qontext is the knowledge layer behind CarOnSale's customer-facing AI assistant. Help-center guidance, FAQs, approved wording and processes for auctions, payments, transport, pickup and complaints live in the context repository, structured for reliable agent use.
Every customer interaction handled by the assistant is powered by Qontext, which provides the context needed to understand the situation and respond correctly.
Take a buyer asking how to file a complaint after delivery. The assistant needs the correct process, the required evidence and the next step. It calls Qontext, retrieves the relevant procedure and uses it to generate the answer.
Ready for live support, with less maintenance
The assistant retrieves support knowledge in under one second, with zero retrieval failures recorded in production. Relevant FAQs, approved wording and operational procedures are available at the speed needed for live customer conversations.
Qontext keeps CarOnSale's support knowledge structured, current and ready for its assistant to retrieve. Its engineering team can develop the assistant without building and maintaining the retrieval stack or deploying code for each knowledge update.