RevelirQA Reaches 100,000 Scored Support Conversations
Revelir AI's automated quality assurance engine now evaluates tens of thousands of customer support tickets monthly against clients' own policies and standard operating procedures. The Singapore-based company, founded in 2025 by Y Combinator alumnus Rasmus Chow, said RevelirQA has scored more than 100,000 conversations since launch.
The system reviews 100% of incoming tickets. Manual QA teams typically sample only 1% to 5% of conversations, leaving most interactions unexamined.
How RevelirQA Works
RevelirQA reads each conversation in full, retrieves relevant policies through retrieval-augmented generation, and scores the interaction against the customer's QA scorecard. It returns a written evaluation with a complete reasoning trace-recording the model used, the prompt, documents retrieved, and the basis for each score.
This audit trail matters in regulated industries. Fintech and financial services require documented justification for every quality assessment, not black-box outputs.
The engine applies the same scorecard to human agents and AI chatbots, giving customer experience leaders one quality standard across both channels. It also tracks sentiment shifts throughout a conversation, surfacing cases where an issue is closed but the customer remains frustrated.
Current Scale and Deployment
RevelirQA runs in production at Indonesian fintech Xendit and travel platform Tiket.com. The infrastructure is built to handle millions of tickets monthly, though it currently processes tens of thousands.
The system connects to helpdesk platforms including Zendesk and Salesforce through an API. It supports Model Context Protocol integration with Claude, allowing customer experience leaders to query support data in plain English and trace answers back to real conversations.
Revelir AI scores tickets in English, Indonesian, Thai, and Tagalog across high-volume support operations in Southeast Asia.
Why This Matters Now
Customer experience leaders are moving past survey metrics like CSAT and NPS toward direct measurement of every interaction. The AI customer service market is projected to reach $83.8 billion by 2033, according to Grand View Research, and quality assurance remains one of its most labor-intensive functions when done manually.
"Manual QA reviews 1% to 5% of tickets, which leaves most customer interactions unexamined," Rasmus Chow said. "RevelirQA scores every ticket against the customer's own QA scorecard, so a missed policy in the conversations no one reads still gets caught."
Product Details
RevelirQA is sold as a subscription priced on ticket volume and deployment fee. Revelir AI also offers Revelir Insights, an analyst layer that answers plain-language questions about support data, and an AI Support Agent that resolves routine queries.
Learn more about AI for Customer Support and AI Agents & Automation to understand how these tools fit into broader customer service operations.
Your membership also unlocks:
