> For the complete documentation index, see [llms.txt](https://atlas.anywhere.now/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://atlas.anywhere.now/contact-center-configuration/building-your-callflow/via-dialogue-studio/configure-advanced-dialogue-studio-features.md).

# Configure advanced Dialogue Studio features

Review advanced Dialogue Studio feature areas for AI-driven flow behavior, AI routing, and contact center settings access.

Use this page to review the advanced Dialogue Studio feature areas that add AI-driven behavior, AI-assisted routing, or reusable contact center settings data to a flow.

## Before you begin

* Confirm that the target Dialogue Studio flow already runs correctly for its baseline routing path.
* Confirm that you can open the target **Service**, **Contact Center**, and flow in the current environment.
* Confirm that any required external AI service, credential, or configuration owner is available before you add advanced nodes to the production design.
* Plan to validate advanced-node behavior in a controlled test flow before you apply the same pattern to production traffic.

{% hint style="warning" %}
Feature names and configuration labels can vary by environment or release. Confirm which advanced Dialogue Studio capabilities are available in your environment before you depend on them in production.
{% endhint %}

## Use AI nodes for AI-driven flow behavior

Use AI nodes when the flow needs a model-driven response, classification result, or decision before it continues to the next branch. The [AI nodes reference](/contact-center-configuration/building-your-callflow/via-dialogue-studio/ai-nodes.md) explains how to configure Azure OpenAI, select each node's input source, and handle its output and failure cases. Do not hardcode API keys in a flow.

### Keep AI-node behavior easy to test

When you add AI nodes:

* keep the prompt or instruction set focused on one decision at a time
* define a deterministic fallback route for unexpected output
* inspect the returned payload in a test path without exposing credentials or sensitive customer content
* confirm whether the result should continue the same conversation, route to another skill, or hand the session to an agent

## Use AI-assisted pathfinding for intent-based routing

Use [AI Pathfinder](/contact-center-configuration/building-your-callflow/via-dialogue-studio/ai-nodes.md#ai-pathfinder) when the next route depends on the conversation context instead of a fixed menu choice. Configure CC Settings Ingest, enable the skills Pathfinder may select, and provide meaningful skill descriptions. Handle an unmatched `Other` result and node errors separately.

### Validate the fallback path

Test these conditions before go-live:

* no eligible route is available
* required transcription, prompt, or route definition is missing
* the AI result does not map cleanly to one configured route
* the external AI request fails or times out

## Use contact center settings nodes carefully

Some advanced Dialogue Studio designs read contact center configuration data from a shared settings source instead of repeating the same values across multiple node chains.

Keep the pattern separated into setup and usage:

* one configuration path exposes the settings data
* one node family ingests or refreshes the settings data
* another node family reads the stored values inside the flow
* cache and collection behavior determine what the flow can retrieve at runtime

### Keep data ownership visible

Use this pattern only when the configuration owner, data source, and refresh expectations are clear.

Good practice is to:

* document which settings collection the flow depends on
* keep item identifiers stable across environments
* confirm how the flow behaves when a requested setting is missing
* retest the cache or refresh path after any configuration change

## Verification

1. Confirm the target environment exposes the advanced node family that the flow design requires.
2. Test one success path and one fallback path for each AI-driven branch.
3. Confirm the flow still reaches a safe next step when the AI response is empty, delayed, or unusable.
4. Confirm any settings-driven node reads the expected collection or item values in the target environment.
5. Confirm that the node settings and runtime behavior match the production flow design in your deployed version.

## Related topics

Go to **Troubleshoot Dialogue Studio** for recurring runtime errors related to connection ownership, message payloads, and transcription setup.

{% content-ref url="/pages/1DZ33M6NXFEl7w5wpxlG" %}
[Troubleshoot Dialogue Studio](/contact-center-configuration/building-your-callflow/via-dialogue-studio/troubleshoot-dialogue-studio.md)
{% endcontent-ref %}

Go to **Use AI nodes in Dialogue Studio** for the configuration procedure and individual node reference.

{% content-ref url="/pages/GlCSD1yum3Por3TyDayO" %}
[Use AI-driven nodes in Dialogue Studio](/contact-center-configuration/building-your-callflow/via-dialogue-studio/ai-nodes.md)
{% endcontent-ref %}

Go to **Integrate Copilot Studio with Dialogue Studio** for an example of an external AI-driven integration pattern.

{% content-ref url="/spaces/iTek8gfJMjuAJhfqvtuU/pages/LwwgOPNpYktaNvgU9Ayg" %}
[Integrate Copilot Studio with Dialogue Studio](/channels-integrations/dialogue-studio-integrations/integrate-copilot-studio-with-dialogue-studio.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

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

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://atlas.anywhere.now/contact-center-configuration/building-your-callflow/via-dialogue-studio/configure-advanced-dialogue-studio-features.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

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.