AI Agents & Chatbotsfree

Evaluation metric example: Correctness (judged by AI)

by David RobertsUpdated Aug 2026

RequiresOpenAI

1/5

STEPS · 8

Starts on a Evaluation event

On chat or evaluation input, runs an AI agent and scores answer correctness using OpenAI as the judge.

AI Email Agent

On a chat message, runs an OpenAI agent that reads Google Sheets and sends emails via Gmail.

On a Telegram message, runs an OpenAI agent with conversation memory, logs the exchange to Google Sheets, and replies in Telegram.

On a Telegram message, an OpenAI agent calls external APIs through HTTP tools and replies via Telegram with conversation memory.