The first section of this page highlights three key metrics that provide insights into the chatbot's performance over the selected time frame. By default, the dashboard presents data from the two preceding weeks, excluding the current week.

All Conversations: The total number of conversations that were closed 15 minutes after the last customer interaction.
Deflected Conversations: Conversations that did not involve a handover request but reached at least one deflection tag.Example: A conversation triggers six use cases, including one misunderstanding, and reaches a deflection tag twice without a handover request. This conversation is classified as deflected.
Handovered Conversations: Conversations that included at least one handover request.Example: A conversation triggers six use cases, including one misunderstanding, reaches a deflection tag twice, and includes a handover request once. This conversation is classified as handovered.
Comparison to Previous Period: The previous period is calculated as the same number of days ending just before the start of the selected time frame.
Conversations Over Time: Conversations are categorized into deflected, handovered, and others (e.g., misunderstood and neutral conversations). This data can be viewed daily, weekly, or monthly, depending on the selected time frame.

Conversations Breakdown

This section provides a detailed analysis of your bot’s conversations, helping you understand how customers are interacting with it.

Conversation Types

Conversations are categorized based on the following rule-based aggregation system:

Handovered Conversations: Conversations that include at least one handover request.
Example: A conversation triggers six use cases, including one misunderstanding, reaches a deflection tag twice, and includes a handover request once. This conversation is classified as handovered.

Ticket Handovered Conversations: Handovered conversations that include at least one ticket request.Example: A conversation includes one ticket handover request and one live chat handover request. This conversation is classified as a ticket handover.
Live Chat Handovered Conversations: Handovered conversations without a ticket request but with at least one live chat request initiated from a Flow Builder use case.
Advanced Builder Handovered Conversations: Handovered conversations without a ticket request but with at least one live chat request that is not initiated from a Flow Builder use case.Note: Handover request types in the Flow Builder are based on the step reached in a use case (e.g., ticket or live chat handover steps). Handover requests initiated from a use case set up in the Advanced Builder are exclusively live chat handovers. Handovers managed via API calls in the Advanced Builder are not tracked in analytics.
Automated Conversations: Conversations that do not include any handover requests.
- Deflected Conversations: Automated conversations that reached at least one deflection tag._Example: A conversation triggers six use cases, including one misunderstanding, reaches a deflection tag twice, but does not include a handover request. This conversation is classified as deflected.

Action Deflected Conversations: Deflected conversations that reached at least one action deflection tag.Example: A conversation reaches one action deflection tag, one information deflection tag, and one deflection tag in the Advanced Builder. This conversation is classified as action deflected.
Information Deflected Conversations: Deflected conversations that did not reach any action deflection tag but reached at least one information deflection tag.Example: A conversation does not reach any action deflection tag but reaches one information deflection tag and one in the Advanced Builder. This conversation is classified as information deflected.
Advanced Builder Deflected Conversations: Deflected conversations that did not reach any action or information deflection tags but reached at least one deflection tag in the Advanced Builder.Example: A conversation does not reach any action or information deflection tags but reaches one in the Advanced Builder. This conversation is classified as advanced builder deflected.Note: Deflection types in the Flow Builder are based on the advanced settings of the answer step marked as deflected (e.g., action vs. information types). Deflections triggered from a use case set up in the Advanced Builder are classified as an Advanced Builder type by default.
Misunderstood Conversations: Conversations that did not reach a handover request or a deflection tag but triggered at least one misunderstanding use case.Example: A conversation triggers six use cases, including one misunderstanding, but does not reach a handover request or a deflection tag. This conversation is classified as misunderstood.Note: By default, your bot is configured with a Misunderstanding/Out-of-Scope use case in the Advanced Builder. If the NLP model cannot match a customer request with an existing intent, it will automatically trigger the Misunderstanding use case, prompting the customer to rephrase their request.
Neutral Conversations: Conversations that did not reach a handover request, a deflection tag, or trigger a misunderstanding use case.Example: A conversation triggers six use cases, none of which are a misunderstanding, and does not reach a handover request or a deflection tag. This conversation is classified as neutral.

Quantifiers

Volume of a Conversation Type: The total number of conversations recorded for that specific type.
Rate of a Conversation Type: The percentage of conversations of a specific type out of the total conversations recorded.
Progress on Rates: Progress is shown in percentage points (pp), representing the difference in the rate of a specific conversation type between the current and previous periods.
Example: If the deflection rate was 51.4% in the current period and 50.8% in the previous period, the progress is +0.6pp, calculated as (current period - previous period).

The cards focusing on deflection and handover details highlight the following elements:

Highest Volume: Quickly identify the conversation type with the highest volume within each group (i.e., deflected or handovered conversations). The top-ranking type is displayed along with its rate progress.
Ranking: The gauge graph shows the distribution of each type within its group, sorted by volume. The absolute volume of each type is displayed in the tooltip when hovering over the gauge graph.

Languages and Channels

This section provides insights into the languages and channels used by your customers during conversations. The top five user channels are ranked by the volume of conversations recorded.

The cards focusing on languages and channel types highlight the following elements:

Language: The language used by the bot during a conversation.
Channel Type: The type of channel through which customers engaged with the bot.
Highest Volume: Quickly identify the language or channel with the highest volume of conversations during the selected time frame. The top-ranking variable is displayed along with its rate progress.
Ranking: The gauge graph shows the distribution of the top three variables (languages or channels) used during conversations in the selected time frame. The variables are sorted by volume, and their absolute values are displayed in the tooltip when hovering over the gauge graph.