Custom Instructions

Sometimes, you want to pass in user information for the AI assistant to consider or provide instructions for the AI assistant to follow.

When using the @inkeep/cxkit-react components, you can pass this information using prompts as part of the aiChatSettings configuration.

You can make these prompts dynamic based on user specific information (e.g. for logged in users) or based on the information from your CRM, application logic, or other sources.

prompts apply to the entire chat session, not just the first message.

Add user context

Contextual prompts can be used to provide additional details about the user that might be helpful information for the AI assistant to consider while answering.

Often, this is used to dynamically provide information about the user that can be inferred by their account (if authenticated to your application) or based on the page they are currently viewing.

Examples

For providing context based on the user:

`The user is currently viewing documentation page ${url}.`

"The user is new to the platform." // (for user life cycle stages)

For providing context based on a product:

`The question is about ${product}.` // (for teams with multiple products)

"The user is using the cloud hosted version of the application." // (for teams with multiple versions or pricing tiers)

Add custom rules

Prompts can be used to customize or guide the behavior of the AI assistant using instructions in natural langauge.

By default, the AI assistanth has sensible defaults for staying on topic, brand-protection, faithfulness to the source content, and other guardrails.

However, prompts can be useful to further customizing tone, having the assistant ask more clarifying questions, adding topics it should refuse to answer, or otherwise customizing the exact experience to your scenario.

See Rules & Prompts for our recommendations and examples around how to prompt LLMs.