In the news
Introducing LangSmith Tuned Evaluators
LangChain · Published · 3 min read
In 30 seconds
- What happened
- LangChain released Tuned Evaluators for LangSmith, starting with Perceived Error detection that flags agent mistakes in production conversations automatically.
- Why it matters
- Teams running conversational agents in production need continuous quality monitoring without manually sampling traces or managing expensive frontier model evaluations.
- Watch out
- Perceived Error evaluator requires threads with at least two human-AI message pairs and evaluates within 12 hours; currently available only in US on Plus and Enterprise plans.
Listen to this summary
- agent
- eval
The patterns behind this
- Error Handling and Recovery Patterns
- Eval-Driven Development (Agent CI)
- Conversational Interface Patterns
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.