Using Journey Maps with Interaction Analytics


The “Journey Map” feature from Talkdesk Interaction Analytics™ enhances the conversational efficiency of virtual agents by analyzing customer journey flows. 

“Journey Map” allows identifying areas for improvement along the customer journey and detecting issues within Talkdesk Autopilot-initiated conversations. These may include:

Journey breaks such as:


Accessing the Journey Map


Once you have access to this feature, within the Interaction Analytics app, select Journey Map [1].


Applying Filters

By default, the following filters are applied:

  • “When”: The interval period is restricted to the last 7 days.
  • “Channels”: Only Autopilot-initiated conversations are considered.
  • “Virtual Agents”: By default, the top 3 flows with more interactions are presented.


You can also filter by any Autopilot flow you wish. To do so, please follow the instructions below:


1. Go to Filters [1].


2. Select the Virtual Agents [2]. Note that you can only choose up to 3 Virtual Agents simultaneously.

3.  Hit Apply [3]


If you wish to dismiss one or several “Virtual Agents” you have previously selected, click X (see an example below) or individually unselect each of them on the “Filters” side panel.

The filters applied will be preserved on the local storage. If you exit and come back again, the same filters will be applied.


Please note that if the user clears the browser cache, it will reset to default. To clear and go back to the default filters, you can select Reset [1] filters.


Identifying the Autopilot


To identify the Autopilot, you can check its name by hovering your mouse over the first intent detected.


Discovering Flows 


The flows can be expanded by clicking on the intent nodes [1] throughout the journey. 

The complete flow will be progressively unveiled by the user's navigation.


Note: To go back to the default view (all flows retracted), click on “1st intent detected” node.

Identifying Pain Points 


Escalations to Human Agents

Channels are identified by colors (green for "Autopilot Digital" and purple for "Autopilot Voice"). 


Escalations are identified by a purple “persons” icon [1]. By clicking on the purple flow, you will be redirected to the "Search" page, so you can analyze the selected interaction in more detail.


Fallback intents


Used in Autopilot when no applicable intent or entity is captured and the bot responds with a predefined message that expresses that it was unable to understand the customer's needs. These intents are also displayed in the “Journey Map” so that the user can analyze if they are causing customer frustration, in order to make the appropriate changes in the respective Autopilot Flow.


Journey Loops


When we have intents that are intercalated (e.g. intent A followed by intent B, followed by intent A) it may indicate that the customer is struggling because the Autopilot is unable to provide a proper response. In this case, to avoid escalations, users can go to the “Automation Designer” and change the Autopilot flows based on these findings.




Displayed by a red flow. By clicking to see related interactions [1], you will be redirected to the "Search" page, so you can analyze the selected interaction in more detail. Keep in mind that drop-offs do not mean necessarily that the conversation is over or was abandoned. It means that no intent was detected afterward, so we are unable to present the entire flow.

Note: Identical sequential intents are only displayed once. For example, if intent “A” is followed by intent “A” and intent “A”, we only display intent “A” once.

All Articles ""
Please sign in to submit a request.