Starting from June 1, 2024, add-on charges are applicable for using the Conversational search feature in addition to your Plus or Enterprise plans. For more information about the pricing plans, see Pricing plans. For more information about terms, see Terms.
Conversational search
Plus Enterprise IBM Cloud Pak for Data
Use conversational search with the IBM Watson® Discovery search integration or Elasticsearch search integration to help your assistant extract an answer from the highest-ranked query results and return a text response to the user.
When you enable this feature, search results are provided to an IBM watsonx generative AI model that produces a conversational reply to a user's question.
To use conversational search, you must have a Plus or Enterprise plan and pay add-on charges. In addition, by enabling the Conversational search feature, you agree to the pricing plans and the terms for using the conversational search.
The watsonx generative AI model is currently hosted only in the Dallas and Frankfurt regions. By default, assistants in all regions except Frankfurt use the model from the Dallas region. Assistants in the Frankfurt region use the model hosted in the Frankfurt region.
Refer to the following topics to configure the conversational search in your assistant:
- Before you begin
- Enabling conversational search
- Tuning conversational search’s tendency to say “I don’t know"
- Configuring your assistant to use the conversational search
- Testing conversational search
- Streaming response for conversational search
Before you begin
You must configure the search integration to enable the conversational search feature. For more information about configuring IBM Watson® Discovery search integration, see Discovery search integration setup. For more information about configuring Elasticsearch integration, see Elasticsearch search integration setup.
Enabling conversational search
You can enable conversational search to give accurate responses to the customer query. In addition, you can enable citations by putting a citation title, which gives the list of references of the source content from where the assistant pulled the responses. You can see the citation title in between the conversational response and the citations.
To enable conversational search, do the following steps:
-
Go to Search Integration window.
-
Set the Conversational search toggle to
On. -
In the Define the text for the citation title, type
How do we know?.The Define the text for the citation title is enabled only when Conversational search toggle is switched to
On.The web chat integration does not support the citation title feature.
-
In the Tendency to say I don’t know, select the tendency to use. By default,
Less oftenis selected. -
Click Save.
Tuning conversational search’s tendency to say “I don’t know"
You can tune the tendency of your assistant to say “I don’t know” in conversational search by selecting one of the following options in the Tendency to say I don’t know section:
RarelyLess oftenMore oftenMost often
When your assistant generates a conversational search response, it evaluates the response and calculates a response confidence score. The assistant compares the response confidence score against your selection in the Tendency to say I don’t know section. If the response confidence score is comparatively high, the assistant responds to the user query with the generated response. If the response confidence score is comparatively low, the assistant does one of the following actions:
- Responds with an “I don’t know” message.
- Falls back to the “No matches” action per the Search routing configuration in your assistant.
'Table 1. Tendency to say “I don’t know” options' shows the options available in the Tendency to say “I don’t know” section.
| Tendency to say “I don’t know” | Response confidence threshold | Assistant behavior |
|---|---|---|
Rarely |
Lowest | The assistant rarely says “I don’t know” because it compares the response confidence score against the lowest threshold. Therefore, your assistant gives a generated response to the end user almost all the time. However, the assistant most likely presents inaccurate or irrelevant responses to the end user. |
Less often |
Lower | The assistant says “I don’t know” more often than the Rarely option. |
More often |
Higher | The assistant says “I don’t know” more often than the Less often option. |
Most often |
Highest | The assistant says “I don’t know” more often than the More often option because it compares the response confidence score against the highest threshold. However, the assistant most likely presents accurate or relevant responses
to the end user. In addition, the assistant presents fewer generated responses to the users and more often responds with an “I don’t know” message or falls back to the “No matches” action. |
Configuring your assistant to use the conversational search
After you enable Conversational search, you must configure the Search routing setting to route your assistant responses to conversational search when no action matches the user response. For more information about the Search routing configuration, see the Configuring the search routing when no action matches topic. To configure your assistant to route to conversational search for specific topics or actions, you can add search as a step in a new or existing action.
When your assistant receives no search results from Elasticsearch or Discovery in response to a user query or when its connection to Elasticsearch or Discovery fails, your assistant responds to the user with a failure message. You can configure the failure messages for no search results and a failed connection in the Search integration settings.
Testing conversational search
You can test the conversational search in actions preview, the preview page, or by using the preview link.
In this example, the user asks, Tell me about a custom extension. Search results are pulled from your knowledge base when the conversational search is Off. In this case, the answer is returned as a list of cards that
are relevant to custom extensions.
When the conversational search is On, the same search results are pulled from your knowledge base. The results are passed to an IBM watsonx generative AI model. This model produces a conversational reply to the user's question,
in the form of a text response about custom extensions.
The web chat integration does not support the citation title feature. Hence, the preceding image does not display the default text in the citation title.
Streaming response for conversational search
Streaming response for the conversational search uses watsonx.ai capabilities to provide continuous, real-time responses in your assistant. By default, the streaming response is disabled for the web chat and the assistant preview panels.
By using the streaming response support feature, you can reduce the wait time for the response.
To enable streaming response, do the following:
- Go to Home > Preview > Customize web chat.
- Click the Styles tab.
- Set the Streaming toggle button to
On. - Click Save and exit.