Bot Ratings
Bot rating functionality allows you to record feedback and performance metrics for your conversational AI agents. The upvote and downvote operations provide a structured way to track bot quality, identify areas for improvement, and measure user satisfaction over time.
The rating system helps you gather both quantitative and qualitative feedback about bot performance. Numerical values provide metrics for tracking improvements over time, while optional reason text captures specific details about what worked well or needs improvement. This data is invaluable for iterating on bot configuration, refining backstories, and optimizing overall performance.
Upvoting Bots
Upvoting a bot records positive feedback with a configurable value between 1 and 100, with 100 being the default maximum positive rating. You can optionally include a reason to document why the bot performed well, which helps with analyzing patterns in successful interactions and understanding what aspects of the bot's behavior are most effective.
Ratings are associated with your user account and the specific bot being evaluated. The system includes rate limiting to prevent abuse and ensure that ratings reflect genuine feedback rather than artificial manipulation. If you exceed the rate limit, the operation completes successfully but the rating is not recorded.
Use bot ratings to establish quality baselines, identify bots that need attention, compare performance across different configurations, and make data-driven decisions about bot improvements. The combination of numerical scores and contextual reasons provides comprehensive insight into bot effectiveness.
Downvoting Bots
Downvoting complements the upvote functionality by recording negative feedback and identifying problematic bot interactions. This helps you quickly spot issues, track patterns in bot failures, and prioritize areas that need improvement.
Downvote values range from -100 to -1, where -100 represents the most severe negative feedback. Including a reason with your downvote provides critical context about what went wrong, enabling targeted improvements to bot configuration, backstory, or connected resources.
Analyzing downvotes helps identify common failure patterns such as inaccurate information, inappropriate responses, inability to handle certain questions, or violations of intended behavior. Use this feedback to refine your bot's backstory, adjust connected datasets, or implement additional safeguards through moderation settings.
Like upvotes, downvotes are subject to rate limiting to ensure data integrity. The combination of upvote and downvote data provides a complete picture of bot performance, enabling you to calculate overall satisfaction scores, identify trends, and make informed decisions about bot optimization priorities.