# Add New Prewritten Responses

## Before Getting Started <a href="#our-recommendations" id="our-recommendations"></a>

{% hint style="danger" %}
Input Training is **VERY** important. The quality and breadth of your inputs will determine how well training works for your Prewritten Response.&#x20;
{% endhint %}

{% hint style="danger" %}
Always double-check that you don't already have an existing response for a specific topic before creating a new one! To check, you can do the following:

* Filter the response library to "All" and search topic keywords&#x20;
* Filter the response library to "Response Name" and search topic keywords&#x20;
* Filter the response library to "Content" and search topic keywords&#x20;
  {% endhint %}

## Add a New Prewritten Response <a href="#add-a-new-prewritten-response" id="add-a-new-prewritten-response"></a>

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FsxN9AGLZMIc5C8dx3zU7%2Fuploads%2FmbYJp4mgcC9tppIDUhcy%2Fadd%20prewritten%20v3.mp4?alt=media&token=077c72ab-10cf-411a-94f6-d442c37e9360>" fullWidth="true" %}
How to Add a New Prewritten Response
{% endembed %}

1. In the Satisfi Labs dashboard, go to **Studio -> NLP Manager -> Responses**
2. Click "+" in the top right of the Responses Dashboard.
3. Ensure that your new response has the following fields completed:

<table><thead><tr><th width="424">Field</th><th>Details</th></tr></thead><tbody><tr><td>Volume Name</td><td>Your Volume Name.</td></tr><tr><td>Channel Name</td><td>If the volume name is not selected first, no channels will appear. The typical default channel is your Content Channel.</td></tr><tr><td>Language</td><td>English is the only supported language at this time.</td></tr><tr><td>Response Name</td><td>Response Names should be all lowercase, with no spaces or special characters except for underscores. Example: add_new_user </td></tr><tr><td>Content Group</td><td>The default group is <strong>General</strong>; however, you can use this field to help organize your responses in whatever works best for you.</td></tr><tr><td>Content Type</td><td>Select <strong>Prewritten Plus</strong></td></tr></tbody></table>

{% hint style="info" %}
![](/files/1jItY0e0MHLx4nnXvbRm) &#x20;

Speed up the process by cloning an existing prewritten plus response! Simply click the clone icon and update the response name, text, and enrichments.&#x20;
{% endhint %}

4. Add the response including any relevant buttons, quick replies, or rich media.
5. Hit **Save New**
6. Select the toggle labeled **Input Training** and begin adding sample inputs. These should be questions end users would ask to trigger this response.&#x20;

{% tabs %}
{% tab title="Number of Sample Inputs" %}
We recommend having a total of **10-15 inputs** per Prewritten Response. You must manually write at least 3 chat-like inputs before generating additional inputs; however, if users prefer, you can manually write out all inputs instead of using the Generate Inputs feature.
{% endtab %}

{% tab title="Length of Sample Inputs" %}
One-word inputs will be treated **ONLY** as a keyword exact match. Input Training functionality uses a semantic similarity search, which means full sentences/questions for the inputs work best.

**Example Topic:** Users are looking for more information about the Satisfi Labs Conference

<table><thead><tr><th width="415">Good Sample Inputs 🤩</th><th>Why We Love It</th></tr></thead><tbody><tr><td>I’d like to learn about the Satisfi Labs Conference</td><td>Full Statement</td></tr><tr><td>When is the the Satisfi Labs Conference?</td><td>Full Question with ‘?’</td></tr><tr><td>How do I get tickets for the Satisfi Labs Conference? / Tell me about the admission cost for the Satisfi Labs Conference</td><td>Variety in Manually Written Inputs</td></tr></tbody></table>

| Bad Sample Inputs 😕                                                                                                | Where Improvement Can Be Made                   |
| ------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------- |
| Satisfi Conference                                                                                                  | Only Two Words/Not a Full Statement or Question |
| How to buy Satisfi Labs Conference tickets? / How to buy tickets for Satisfi Labs Conference? / How to buy tickets? | Lack of Variety in Manually Written Inputs      |
| {% endtab %}                                                                                                        |                                                 |

{% tab title="Helpful Hints" %}
Learn more about the Input Director tool and helpful hints here: [Input Director](/resource-center/ai-agent-engine/manage-your-content/input-director.md#what-is-the-input-training-tool)
{% endtab %}
{% endtabs %}

7. Select the **Generate Inputs** button to add additional training.
8. Delete any undesired sample inputs by selecting the toggle to the left of the input and clicking **Delete Selected Inputs**.
9. Press **Publish** and test the new response.

{% hint style="warning" %}
In certain prewritten responses, you might find it necessary to suppress the thumbs up/down feature. Please review the instructions here to learn more: [Suppress Thumbs Up/Down](/resource-center/ai-agent-engine/manage-your-content/best-practices/suppress-thumbs-up-down.md)
{% endhint %}


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.satisfilabs.com/resource-center/ai-agent-engine/manage-your-content/response-types/prewritten-plus-responses/add-new-prewritten-responses.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.