Agnost AI Launches: Product Analytics for Chat and Voice Agents

Agnost AI, founded by YC S26 founders Shubham and Parth, is a product analytics platform specifically for teams building chat and voice agents. It ingests production conversation messages via SDKs or OpenTelemetry, then surfaces behavioral failures that traditional evals miss — like users rageprompting (cursing at the agent), repeatedly rephrasing the same request, correcting the agent, asking for missing features, or leaving after a technically successful answer.

How it works

Teams send agent conversation messages through SDKs (available on PyPI and npm) or OTel, optionally with metadata like account, plan, source, organization. Agnost AI clusters conversations into product-specific intents. Feature requests and bugs are default categories; most other clusters are created dynamically from the data and evolve over time. You can create your own cluster in plain English. If a cluster gets too broad, the system splits it. If a new pattern appears, it suggests it.

Real example

One AI video editor company used Agnost AI to find hidden feature requests: around 70 users wanted auto-subtitles, expressed as “add this text in this frame” (12x in a single session), “can you caption it,” “give me transcript of audio,” and variations across languages. The team later built the feature.

Technical approach

The hard part was doing this over millions of messages without sending everything to an LLM. In ClickHouse, “fetch the last 50 events by time across conversations” and “fetch all events in this conversation” require different sort orders, so they iterated on sorting keys, partitions, materialized views, and projections. For finding new clusters, sending everything through an LLM was too slow and expensive. They split conversations into segments based on cosine drift, run BIRCH to compress the candidate space, and then use HDBSCAN-like clustering on the smaller set. For matching existing clusters, they use embeddings, smaller classifiers/BERT-style models, and LLMs only as fallback for ambiguous cases.

Pricing

• Starter: Free — up to 1,000 messages/month, 7-day retention
• Pro: $499/month — up to 100,000 messages/month, 90-day retention, priority support
• Enterprise: Custom — higher volume, security, retention, audit logs, custom SLAs

Status

Agnost AI is live with multiple companies, ingesting ~1M chat and voice messages per day. They are SOC 2 Type 1 compliant (Type 2 in progress). Each customer's data is used only for that customer. An interactive demo with no signup is available at app.agnost.ai.

Detect these signals today: rageprompting, repeated retries, setup friction, churn risk, and more — directly from real production conversations.