AIOptions interface

Optional. If true, the AI system will allow the planner to loop.

allow_looping?: boolean

Property Value

Remarks

The default value is true.

Looping is needed for augmentations like functions and monologue where the LLM needs to see the result of the last action that was performed. The AI system will attempt to autodetect if it needs to loop so you generally don't need to worry about this setting.

If you're using an augmentation like sequence you can set this to false to guard against any accidental looping.

Optional. If true, the AI system will enable the feedback loop in Teams that allows a user to give thumbs up or down to a response. Default is false.

enable_feedback_loop?: boolean

Property Value

Remarks

At this time, there is no activity handler support in the Teams AI Library to handle when a user gives feedback. To make use of the feedback loop, use the app.feedbackLoop route registration. https://github.com/microsoft/teams-ai/blob/main/getting-started/CONCEPTS/POWERED-BY-AI.md

Optional. Maximum number of actions to execute in a single turn.

max_actions?: number

Property Value

Remarks

The default value is 25.

Optional. Maximum amount of time to spend executing a single turn in milliseconds.

max_time?: number

Property Value

Remarks

The default value is 300000 or 5 minutes.

Optional. The moderator to use for moderating input passed to the model and the output returned by the model.

moderator?: Moderator<TState>

Property Value

The planner to use for generating plans.

planner: Planner<TState>

Property Value