Response Settings

​

Response Length

Control the average length of a response presented to a user. There are four different options to choose from.

Intelligent - The response length will depend on the nature of the inbound interaction. Answers that require more depth will be longer. Interactions that are naturally shorter will be concise. We recommend using this response length option to start.

Concise - Responses will be brief and focused on essential information.

Explanatory - Responses will be in-depth, exploring topics in detail.

Custom - Set your desired response length in the number of words.

​

Creativity

Creativity refers to your Digital Mind’s ability to reason based on information available on the internet.

Strict - It will only discuss topics it’s trained on.

Adaptive - It will infer what you might say in new situations based solely on your training data.

Creative - Utilizing information from the internet when generating responses. High creativity is beneficial for giving personalized advice.

​

Dynamic Questions

If you want your Digital Mind to ask follow up questions at the end of a response, toggle on Dynamic Questions. You can provide specific guidance about questioning style (e.g., formal vs. casual, brief vs. detailed) in your Custom Instructions.

​

Additional Configurations

Personalize Responses – By toggling this on, previous conversations from users will be remembered for personalized responses. It will call them by name and be able to tie previous conversations to current ones.

Recency Bias – By turning this on, your Digital Mind will prioritize recent information that it is trained on. For example, if you upload a health blog that has been in existence for 10 years, the content from 9 years ago might not be as relevant. Activating recency bias will prompt your Digital Mind to reference recent content first when generating responses.

Favor New Content – This setting pertains to the user experience. By turning it on, users will be presented with fresh content they have not interacted with recently. The trade-off to utilizing this feature is that certain responses may not be given consistently. If a user asks multiple questions about a specific topic, it will seek new pieces of information from training to present.