> For the complete documentation index, see [llms.txt](https://docs.aisera.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.aisera.com/aiseragpt/how-to-set-up-an-aiseragpt-bot/view-conversational-ai-policy.md).

# View Conversational AI Policy

If your bot uses the Conversational AI 2.0 option, the **View Policy** button is now called the View Conversational AI Policy button. The parameter on a 1.0 bot that was called, **Agent Persona Snippet** has been renamed to  **Agent Response Format Instructions** and allows you to specify instructions for the 2.0 bot response.

<div align="left"><figure><img src="/files/rqMBPW8t03YIzk6gaRAa" alt="" width="217"><figcaption><p>Conversational AI 2.0 View Policy</p></figcaption></figure></div>

The Aisera **Conversational AI** **2.0** bots support an app policy variable where custom instructions can be entered to fine tune the generated LLM output.&#x20;

These instructions can be appended to the standard generation prompt. This way, the change will not impact other bot behavior:

* It will not be lost if the app is switched to a different LLM
* It will not inadvertently impact other features or data
* Any product updates to the prompt will be applied correctly to the app (regardless of the customization)

### Agent Response Format

This variable allows you to provide  context and tone for the conversation.\
\
**Example App Policy** - Please keep the tone formal. Never use bold or italics formatting. If the context includes any instructions, provide your response in a step-by-step manner using bullet points.

<div align="left"><figure><img src="/files/TL5BckKxyX4gl6UAydQf" alt="" width="435"><figcaption></figcaption></figure></div>

### Show Images in KB Output

If your **Knowledge Base Articles** have images embedded in the content, then those images may be used in generated RAG responses that are sent to users on Webchat, Teams, or Slack channels.

<div align="left"><figure><img src="https://docs.aisera.com/~gitbook/image?url=https%3A%2F%2F2983236984-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FiZkLJr3EjXkd2tHYiQJP%252Fuploads%252Fk3rAAcOksFIlrgW27FXl%252F2.png%3Falt%3Dmedia&#x26;width=768&#x26;dpr=4&#x26;quality=100&#x26;sign=bbe5f5f8&#x26;sv=2" alt="" width="563"><figcaption></figcaption></figure></div>

If you do not wish to use images in generated answers, you can turn **OFF** this feature by unchecking the option. You might choose not to use images in your fulfillment answers if you find that images are not relevant to your application/bot users or if the image size changes the formatting in unexpected ways.

In addition to the **Agent Response Format** and **Show Images in KB Output,** you can set your application/bot to work with ontologies, detect user countries, address users with profile information, use strict validation, and show images from KB Articles when relevant.