After completing the prerequisite steps, you should have an Airlines knowledge base powered by the LP_Airline domain intents, as well as two bots that are connected to agents. You can now move your focus to Conversation Assist to configure its capabilities. This involves defining the rules that determine when the answers and bots are offered to agents.

Step 6: Configure the bots for Conversation Assist

  1. Open the menu on the left side of the page, and select Automate > Conversation Assist to open Conversation Assist.

    The Automate menu with the Conversation Assist menu option highlighted

    If it's your first time accessing Conversation Assist, you might see a window that introduces you to the application. Just click Get started in the window, and then click Got it at the next prompt. You'll be on the Recommendation Sources page at this point, which is good.

  2. If necessary, click Recommendation Sources at the top of the page, and then select the Bots tab.

  3. Click + Add rule in the upper-right corner.

  4. Under General, specify the following:

    • Name: Recommend Seating/Booking bots
    • Description: Recommends the Seating bot and Booking bot to agents assigned to the Agent skill.

    Specifing a name and description for the bot rule

    Recommendations are tied to skills, so next you must specify the skills that you want this rule to apply to.

    A skill can be used in only one bot rule.

  5. Still under General, for Skills, select "Agent."

    Selecting Agent as the skill for the rule

    Earlier, you assigned an "Agent" skill to the logged-in human agent. You want to recommend the bots to agents assigned to the Agent skill, so you must specify "Agent" in the rule's Skills field.

  6. Scroll down to the ADD-ONS section.

    A rule add-on completes the rule’s definition. You must define at least one rule add-on because, at a minimum, that’s where you specify the bots to use in the rule.

    If you define multiple add-ons, the order of the add-ons matters: At runtime, the add-ons are evaluated in order, and the first one that’s matched is executed. But in this rule, we need and use only one add-on.

  7. Within the first add-on, select both the "Booking Bot" and the "Seating Bot." Then click Save and activate.

    Selecting Agent as the skill for the rule

    In an add-on, you can further narrow the scope of who receives the recommendations based on agent groups and/or profiles. We've kept things simpler for this tutorial—i.e., just skill-based—but know that it's possible in both bot rules and knowledge base rules.

  8. Click the Back button in the upper-left corner of the Rule Details page.

    The Back button

Step 7: Configure the knowledge base for Conversation Assist

  1. Still on the Recommendation Sources page, select the Knowledge Bases tab.

    The Knowledge Bases tab under Recommendation Sources

  2. Click + Add rule in the upper-right corner.

  3. Under General, specify the following:

    • Name: Recommend Airline FAQs
    • Description: Recommends answers from the Airline FAQs knowledge base to agents assigned to the Agent skill.

    Specifing a name and description for the knowledge base rule

    Here too, recommendations are tied to skills, so next you must specify the skills that you want this rule to apply to.

    A skill can be used in only one knowledge base rule.

  4. Still under General, for Skills, select "Agent."

    Selecting Agent as the skill for the rule

    Similar to what you did in the bot rule, here you want to recommend answers from the Airline FAQ knowledge base to agents assigned to the Agent skill. So you must specify "Agent" in the rule's Skills field.

  5. Scroll down to the ADD-ONS section.

    Here too, a rule add-on completes the rule’s definition. You must define at least one rule add-on because, at a minimum, that’s where you specify the knowledge base to use in the rule.

    If you define multiple add-ons, the order of the add-ons matters: At runtime, the add-ons are evaluated in order, and the first one that’s matched is executed. But in this rule, we need and use only one add-on.

  6. Within the first add-on:

    1. Select the "Airline FAQ" knowledge base.
    2. Turn on the enrich answers via Generative AI toggle.
    3. Click Save and activate.

    Selecting the Airline FAQ knowledge base in the rule

    For the purpose of this tutorial, we're relying on the default Enrichment prompt. We're also relying on the default values for several settings, e.g., the Max # of recommendations to offer at any one time.

Learn about answers enriched via Generative AI (they're smarter, warmer, and better). And learn about setup in Conversation Assist.

Customize your POC: If you built out additional resources in previous steps, make sure to turn on those resources too. Depending on the number of resources, it might be useful to increase the number of recommendations that you display to your agents when they are in a conversation.

Step 8: Create the Web messaging engagement

Updates to the Conversation Assist configuration can take up to 20 minutes to be reflected in conversations. While waiting for these changes to take effect, create a web messaging engagement that connects to the “Agent” skill.

  1. Open the menu on the left side of the page, and select Engage > Campaign Builder to open Campaign Builder.
  2. Click + Add campaign in the lower-left corner.

  3. On the page that appears, specify the following:

    • Campaign name: Enter "Conversation Assist Testing."
    • Campaign goal: Click + Add goal. Select "Interact with consumers." And click Done.
    • Engagement: Click + Add engagement. For the engagement source, select "Web." In the gallery that appears next, select a messaging template, and click Next. In the resulting Engagement Settings, select "Messaging" for the Conversation type. For Routing, select "Specific skill," and then select the "Agent" skill.

    The Engagement settings, showing Routing to a specific skill named Agent

    It is important that this engagement is routing to the "Agent" skill that has been assigned to the human agent. As the Conversation Assist settings have been configured to recommend bots and answers to this particular skill, skipping this important step will result in recommendations not being sent.

  4. Click Next, and continue through the workflow, accepting the default values. When you’ve reached the last option, "Behavioral targeting library," click Done in the lower-right corner.

  5. Back on the "Conversation Assist Testing" page, confirm that the skill being routed to is "Agent." If so, select the orange Publish button in the upper-right corner. Once published, there will be a visible Running icon The Running icon to the right of the engagement.

    The final, published campaign that has an engagement that routes to the Agent skill

What's next?

Continue on to the next step in the tutorial.