Ringg's AI agents resolve up to 65% of customer calls with OpenAI

OpenAI published a customer story on September 23, 2026 about Ringg, a voice and chat agent platform whose agents handle over 7 million connected calls a month and resolve up to 65% of routine inquiries without a human. Its customers average a CSAT score of 4.8.

The startup, which works with large consumer businesses in India, built the platform after seeing support teams struggle with rising call volumes and fragmented manual systems. Moving suitable real-time workloads from GPT-4.1 to GPT-5.6 cut model costs by roughly 90% while preserving quality and latency.

“We need low latency, reliable tool use and economics that work at scale. OpenAI gave us the balance we needed,” said Ringg co-founder Siddharth Tripathi.

Routing each task to the right model

Ringg routes work by task. GPT-4.1 handles most real-time voice and chat traffic, GPT-5.6 Luna takes requests where its speed, quality or price fits better, GPT-5.6 Terra runs post-call analysis such as summaries and sentiment classification, and GPT-5.6 Sol supports evaluation and prompt improvement.

An orchestration layer executes actions across CRMs, ticketing platforms, payment systems and internal APIs, escalating to a human with a summary when needed. Specialized subagents handle qualification, support, verification and escalation while one conversation continues across voice, chat, WhatsApp and web. When context approaches 80,000 tokens, the system compacts it into a structured summary.

Testing models before production

Ringg tests models on historical conversations and simulated flows, then releases them to a small share of traffic. In post-call analysis, GPT-5.6 Terra outperformed Gemini 2.5 Flash and reached up to 97% accuracy on common regional languages.

In production, a router monitors latency and endpoint health across regions and shifts traffic when needed. Migrating selected real-time workloads from GPT-4.1 to GPT-5.6 Luna reduced model costs by approximately 90%.

Customer results and next steps

Policybazaar routes more than 57,000 requests through Ringg and handles 67% of calls without a human; its average response time fell from 8-12 minutes to under 60 seconds. Practo reports 85% first-call resolution, responses under three seconds, 70% lower operating costs and over 1,000 daily appointment bookings. Groww resolves 72% of inbound IPO, futures and options queries via self-service in two minutes.

Ringg is now building browser agents on OpenAI's computer-use capabilities for onboarding, KYC, IT troubleshooting and claims processing, plus a context layer that carries customer details across channels.