This will be a prepackaged Suite of Tests designed to challenge the chatbot adversarially and jailbreak it.

Evaluations are very general and have no context on the chatbot's role, any policy, or in general conversation context for what is deemed quality or good. This would have an optional form users can fill out when creating a Bot and upload policy docs etc.

This will add a prepackaged Suite of Tests for quality assurance that is generated specific to each chatbot's use case.

This will add a webhook integration allowing Test Runs or Suite Runs to be executed via webhook.

Allow users to create and manage tags that they can attach to Suites, Environments, or specific Suite Runs. These tags will allow users to more effectively manage and sort their historical Suite Runs for trends or track changes with models, etc.

This would include metrics such as Recall, Precision, BLEU Score, ROUGE Score, Relevance and Groundness, calculated using "ground truth" from evaluation feedback.

Upon finishing a Scheduled Suite Run, configure a group of emails to receive the generated report. Or optionally, allow for Slack or other integrations.

This will allow users to choose which underlying LLM they want bottest.ai to use within its evaluation engine. Allow the option to select from a wide variety of LLM providers (OpenAI, Anthropic, Meta, Google, Microsoft, etc)