次の方法で共有


View and understand Omnichannel Insights dashboards

Omnichannel for Customer Service offers a suite of capabilities that extend the power of Dynamics 365 Customer Service Enterprise to enable organizations to instantly connect and engage with their customers across digital messaging channels. An additional license is required to access Omnichannel for Customer Service. For more information, see the Dynamics 365 Customer Service pricing overview and Dynamics 365 Customer Service pricing plan pages.

Omnichannel Insights

Note

Microsoft Power BI template reports for Customer Service Analytics dashboards and Omnichannel Insights are deprecated as of November 6, 2023. We recommend that you use the out-of-box Omnichannel historical analytics and Customer Service historical analytics, which don’t need additional Microsoft Power BI licenses. You can visually customize and extend these reports to add additional data sources and metrics through data model customization. For details on the deprecation, go to Deprecations in Customer Service.

The Omnichannel Insights dashboard provides KPIs and trends for supervisors to understand the overall state of the support experience at a glance. It also provides insights on the customer sentiments derived using support-specific machine learning algorithms with an ability to slice through different levels in the organizational hierarchy. Supervisors can rely on the information to improve the overall customer support experience.

The overview dashboard is divided into two sections:

  • Omnichannel Insights dashboard

  • Omnichannel Sentiment Analysis dashboard

Omnichannel Insights dashboard

The following illustration is an example of the Omnichannel Insights for Dynamics 365 dashboard.

Omnichannel for Customer Service dashboard.

The following table describes the conversation KPIs.

KPI Description
Incoming conversations The number of conversations initiated by the customer.
Conversations engaged Offered conversations that are engaged by an agent. Customer-to-agent communication can begin at this point.
Average wait time (mins) The length of time, in minutes, a customer is waiting in a queue.
Average wait time Total length of time (in minutes) / Offered
Abandon Rate The percentage of conversations that aren't engaged by agents.
Average Handle Time (mins) Average length of time an agent takes to complete a conversation with a customer. This time considers the time spent by one or more agents to help the customer.
Transfer rate The percentage of conversations that are transferred to another agent/queue.
Average Customer Sentiment Pulse (CSP) The predicted customer sentiment in a given timeframe for a set queue or agent, which indicates the degree of positive sentiment expressed by customers at the end of their interactions.
Average customer effort time Average length of time a customer takes to contact support and complete a conversation with an agent. Only the conversations engaged by an agent are considered for this metric.

Omnichannel Insights for Dynamics 365

Omnichannel Insights reports provide comprehensive information on how overall support is performing across channels. The reports provide administrators and supervisors with a rich visualization and ability to filter across channels, queues, agents, and date ranges to better understand performance and troubleshoot problem areas.

The reporting structure consists of the following sections:

Conversations and channels

This section provides historical visibility into the overall support operations across various conversations and channels.

Conversation.

Channel report.

The KPIs for conversations and channels are listed in the following table. For conversations, these KPIs are applicable to scenarios in which bots escalate to agents or agents directly handle customer calls. The KPIs for channels represent support operations by each channel to help supervisors easily understand how each support channel is performing, and take appropriate actions to improve the overall support experience for customers.

KPI Description Derivation Measure
Incoming conversations The number of conversations initiated by the customers that can be presented to agents. All conversations are considered. FactConversation[InComingConversationCount]
Conversations engaged Offered conversations that are engaged by an agent. Customer-to-agent communication can begin at this point. All conversations are considered. FactConversation[Engaged]
Abandon rate The percentage of conversations that aren't engaged by agents. All conversations are considered. FactConversation[AbandonedRate]
Transfer rate The percentage of conversations that are transferred to another agent or queue. Conversations engaged and conversations that are in the closed state are considered. FactConversation[TransferRate]
Conversation active time Cumulative session active time for a conversation. Conversations engaged and conversations that are in the closed state are considered. FactConversation[ConversationActiveTime]
Conversation inactive time Cumulative session inactive time for a conversation. Conversations engaged and conversations that are in the closed state are considered. FactConversation[ConversationInactiveTime]
Conversation wrap time Cumulative time from the conversation wrap-up start time until the conversation close time. Conversations engaged and conversations that are in the closed state are considered. FactConversationFirstAgentParticipant[ConversationWrapupTime]
Conversation handle time Cumulative session active time for a conversation. Conversations engaged and conversations that are in the closed state are considered. FactConversation[ConversationActiveTime]
Average Conversation active time Total conversation active time divided by the number of conversations handled. Conversations engaged and conversations that are in the closed state are considered. FactConversation[AvgActiveTime(mins)]
Average Conversation inactive time Total conversation inactive time divided by the number of conversations handled. Conversations engaged and conversations that are in the closed state are considered. FactConversation[AverageConversationInactiveTime]
Average Conversation wrap time Total conversation wrap time divided by the number of conversations handled. Conversations engaged and conversations that are in the closed state are considered. FactConversation[AverageConversationWrapupTime)]
Average Conversation handle time Total Conversation active time divided by the number of conversations handled. Conversations engaged and conversations that are in the closed state are considered. FactConversation[AvgConversationTime]
Average Conversation time Average time from the conversation start to conversation end. Conversations engaged and conversations that are in the closed state are considered. FactConversation[AvgConversationTime]
Average Customer Effort time Average time from the conversation start to the conversation wrap-up start time. Conversations engaged and conversations that are in the closed state are considered. FactConversation[AvgCustomerEffort]
Speed to answer The average time customers have waited in the queue before connecting to an agent. Conversations engaged and conversations that are in the closed state are considered. FactConversation[AvgSpeedtoAnswer]
SLA - Speed to answer Number of SLAs met divided by the number of conversations handled. For example, if the Speed to answer is less than 180 seconds, then this is considered met. Otherwise, it isn't met. Conversations engaged and conversations that are in the closed state are considered. FactMessage[SLASpeedtoAnswer]
Customer wait time The average time customers have waited before connecting to agents. This is similar to “Speed to answer” but includes the time waited on each session within a conversation. Conversations engaged are considered. FactConversationParticipant[AvgWaittime(mins)]
Total consult time The time spent on the consult from when the agent joined to when they left in session participant. Only consult sessions are considered as a denominator.
Average consult time Total consult time divided by the total consult sessions. Only consult sessions are considered as a denominator. FactSessionParticipant[AvgConsultTime]
Total monitor time The time spent on the monitor from when the agent joined to when they left in session participant. Only monitor sessions are considered as a denominator.
Average monitor time The total monitor time divided by the total of monitor sessions. Only monitor sessions are considered as a denominator. FactSessionParticipant[AvgMonitorTime]
Average Customer sentiment pulse (CSP) The predicted customer sentiment in a given timeframe for a set queue or agent, which indicates the degree of positive sentiment expressed by customers at the end of their interactions. Conversations engaged are considered. Average of FactConversationSentiment[msdyn_sentimentpulse]

Queue and agents

The Queue and Agent sections provide historical visibility into how each queue and agent is performing across different channels and queues so that supervisors can take appropriate steps to improve the support experience for the customer.

Queue report.

Agent report.

The KPIs for queues and agents are listed in the following table. Metrics in this section are computed based on the session granularity. Each customer contact is defined as a conversation. Each conversation is defined as a session and can be handled by one or more agents.

KPI Description Derivation Measure
Consult sessions The number of sessions accepted by a user in mode = consult. Only consult sessions are considered as a denominator
Total consult time The time spent on the consult from when the agent joined to when they left in session participant. Only consult sessions are considered as a denominator.
Average consult time The total consult time divided by the total consult sessions. Only consult sessions are considered as a denominator. FactSessionParticipant[AvgConsultTime]
Monitor sessions The number of sessions accepted by a user in mode = monitor. Only monitor sessions are considered as a denominator.
Total monitor time The time spent on the monitor from when the agent joined to when they left in session participant. Only monitor sessions are considered as a denominator.
Average monitor time The total monitor time divided by the total of monitor sessions. Only monitor sessions are considered as a denominator. FactSessionParticipant[AvgMonitorTime]
Sessions engaged # Sessions presented to an agent and accepted by an agent Conversations engaged are considered and all conversation states are considered. FactSession[EngagedByAgentSessionCount]
Session rejection rate The number of sessions presented to an agent and aren't accepted by an agent. Conversations engaged are considered and all conversation state are considered. FactSession[SessionRejectionRate]
Transfer rate The number of sessions transferred by an agent. Conversations engaged are considered and all conversation states are considered. FactSession[QueueTransferRate]
Consult sessions The number of sessions where the agent has participated in consult mode. Conversations engaged are considered and all conversation states are considered. FactSessionParticipant[ConsultSessionCount]
Monitor sessions The number of sessions where the agent has participated in monitor mode. Conversations engaged are considered and all conversation states are considered. FactSessionParticipant[MonitorSessionCount]
Average Session active time The total session active time divided by the number of sessions engaged (primary). Conversations engaged are considered and all conversation states are considered. FactSession[AvgActivetime]
Average Session inactive time the total session inactive time divided by the number of sessions engaged (primary). Conversations engaged are considered and all conversation states are considered. FactConversationFirstAgentParticipant[AverageSessionInactiveTime]
Average Session handle time The total session active time divided by the number of sessions engaged (primary). Conversations engaged are considered and all conversation states are considered. FactSession[AvgSessionHandleTime]
Average Session time The average time from the session start to session end for sessions engaged divided by the sessions engaged (primary). Conversations engaged are considered and all conversation states are considered. FactConversationFirstAgentParticipant[AvgSessionTime]
Average consult time The average time the agent spent on a session in consulting mode. The sum of consult time divided by consult sessions. FactSessionParticipant[AvgConsultTime]
Average monitor time The session level average time the agent spent on a session in monitoring mode. The sum of monitor time divided by monitor sessions. FactSessionParticipant[AvgMonitorTime]
Session active time The time an agent actively spent on a session. Conversations engaged are considered and all conversation states are considered.
Session inactive time. The time an agent isn't actively working on a session. Conversations engaged are considered and all conversation states are considered.
Session handle time the time an agent actively spent on a session. Conversations engaged are considered and all conversation states are considered.
Agent total login time The total time the agent is signed in. Note: This metric isn't sliced in any dimension other than from Date and Agent. This is based on the agent sign in and sign out timestamp, and isn't sliced by any other metrics other than from Date and Agent. FactAgentStatusHistory[AgentTotalLoginTime(hrs)]
Agent Total Login Time (hrs) The time an agent is signed in to the Omnichannel application. This is based on the agent sign in and sign out timestamp, and isn't sliced by any other metrics other than from Date and Agent. FactAgentStatusHistory[AgentTotalLoginTime(hrs)]
Agent Available Duration (hrs) The time an agent is in the available state in the Omnichannel application. This is based on the agent sign in and sign out timestamp, and isn't sliced by any other metrics other than from Date and Agent. FactAgentStatusHistory[AgentAvailableDuration(hrs)]
Agent Busy Duration (hrs) Time an agent in the busy state in Omnichannel application. This is based on the agent sign in and sign out timestamp, and isn't sliced by any other metrics other than from Date and Agent. FactAgentStatusHistory[AgentBusyDuration(hrs))]
Agent Busy DND Duration (hrs) The time an agent is in the Busy DND state in the Omnichannel application. This is based on the agent sign in and sign out timestamp, and isn't sliced by any other metrics other than from Date and Agent. FactAgentStatusHistory[AgentBusyDNDDuration(hrs)]
Agent Away Duration (hrs) The time an agent is in the Away state in the Omnichannel application This is based on the agent sign in and sign out timestamp, and isn't sliced by any other metrics other than from Date and Agent. FactAgentStatusHistory[AgentAwayDuration(hrs))]
Agent Offline Duration (hrs) The time an agent signed out of the Omnichannel application. This is based on the agent sign in and sign out timestamp, and isn't sliced by any other metrics other than from Date and Agent. FactAgentStatusHistory[AgentOfflineDuration(hrs) )]

Bot insights

This section provides historical visibility into how bots are performing to help resolve customer support issues.

Bot insights.

The KPIs for bots are listed in the following table.

KPI Description Derivation Measure
Bot conversations The number of conversations initiated by the customer and handled by a bot. All conversations are considered. FactSession[QueueSessions]
Bot resolution rate  The percentage of conversations that were closed by interacting with a bot out of all conversations handled by a bot.  All conversations are considered. FactSession[BOTResolutionRate]
Bot resolution time (mins) The length of time, in minutes, a customer interacted with a bot before the conversation was closed. All conversations are considered. FactSession[AvgResolutionTime]
Bot escalation rate  The percentage of conversations that are escalated by a bot to a human agent. All conversations are considered. FactSession[BotEscalationRate]
Bot escalation time (mins)  The length of time, in minutes, a customer interacted with a bot before the conversation was escalated to a human agent. All conversations are considered. FactSession[BotEscalationTime]

Omnichannel Sentiment Analysis dashboard

The Omnichannel Sentiment Analysis dashboard provides an overview of important KPIs and trends related to the sentiment analysis of conversations.

Omnichannel Insights – Sentiment Analysis report

The following table provides a detailed view of the Omnichannel Sentiment Analysis report.

KPI Description   
Average Sentiment Pulse  The predicted customer sentiment in a given timeframe for a set queue or agent that indicates the degree of positive sentiment expressed by customers at the end of their interactions. For channel and queue, it provides the overall customer sentiment of the conversation. For agent, it provides the customer sentiment specific to the sessions handled by the agent in the conversation. 
% Positive Sentiment  Count of positive sentiment zone conversations divided by total sessions. 
% Neutral Sentiment   Count of neutral sentiment zone conversations divided by total sessions. 
% Negative Sentiment   Count of negative sentiment zone conversations divided by total sessions. 
Conversations with Sentiment prediction  Count of conversations to predict the customer sentiment metrics.

Sentiment zones

Positive sentiment

Positive sentiment is expressed by the customer writing messages that they're happy, pleased, or positive as a result of the support interaction. Positive sentiment requires the positive words to convey positive feeling beyond simple pleasantries or politeness.

For example, when a customer writes "thank you," that's just being polite—it doesn't necessarily mean they're pleased or happy. However, if a customer writes "I can't thank you enough" or "thank you very much," that portrays clear positive feeling. Other examples of simple pleasantries or politeness that shouldn't be mischaracterized as positive are "Please help" and "Yes."

Negative sentiment

Negative sentiment is expressed by the customer when they're disappointed with the support interaction. These can be cases where the customer is describing a problem and is frustrated or unhappy because of how it's currently affecting them. For something to be scored with negative sentiment, the words need to clearly convey that the user is unhappy, disappointed, or frustrated.

Neutral sentiment

Neutral sentiment is represented when a customer's sentiment was neither positive nor negative. A customer's problem statement isn't to be scored as negative; rather, it should be scored neutral unless it contains words that denote emotion. This is true regardless of how severe the user's problem is. Messages with non-specific pleasantries or politeness are scored as neutral and not as positive.

View and filter reports

You can filter the information presented in the reports by selecting Duration, Channel, Queue, and Agent.

Introduction to Omnichannel Insights dashboard

Configure Omnichannel Insights dashboard

Analyze real-time customer sentiment