FROM VERSION 3.61.0

Overview

The Quality Analysis Dashboards display interactions data, analysed through Quality Analysis features, by processing voice conversations and textual interactions:

• Sentiment Analysis: a percentage score reflecting the overall tone of the conversation, categorized as Positive, Negative, Neutral, or Mixed. This score is determined through analysis of keywords and context.

• Post-Call Analytics: A more granular analysis to detect also categories within the conversation (interruptions, silences, specific keywords) and relevant metrics for an in-depth assessment of interaction quality.

The dashboards allow admins to view the information obtained from the Quality Analysis features. This data will result from the processing of raw Quality Analysis data, structured optimally for clear and meaningful visualization via Grafana.

You can access to the available dashboards directly from your Motion server under Dashboard menu:

• Sentiment, with widgets to display sentiment data from voice (inbound and outbound calls and Cally Square projects) and textual channels (Chat, SMS, WhatsApp, Openchannel)

• Post Call Analytics, with widgets to display post-call data, only for voice channel (calls and Cally Square projects)

Requirements

Video

Process Flow in 5 steps

1. User Request: Motion user submits a Quality Analysis request on an interaction present in the system

2. Cloud provider analysis: the conversation is sent to cloud provider for analysis

3. Saving Results: the results are saved in Motion MySQL while the QA Microservice is notified of the new QA results via Bull-MQ (the microservice processes the results and saves them in QA MongoDB)

4. User Access: the Motion user accesses to Grafana (if he has the right permission to access, thanks to JWT token validation).
   Admin can enter in Grafana and choose the relative dashboard or he can click the link of the desired one directly from the Motion Dashboard section, by choosing between Sentiment and Post-Call Analytics dashboard

5. Navigating to Grafana Dashboard: the user navigates to the Grafana Dashboards – at the moment only admin can access to Grafana. If you have knowledge of Grafana, you can also add new widgets, creating customized dashboards.

Dashboards

Quality Analysis - Sentiment

On this dashboard, with different widgets and filters, you can view the sentiment analysis of all interactions, related to voice channel (voice calls and Cally Square) and digital channels (except mail)

Open

You can filter by:

• Cloud provider: AWS or OpenAI.

If you launch multiple sentiment analysis, you will view different records. In fact, every time you launch a new sentiment (with the same cloud provider or with the other one), it is considered as new data.

• direction of interaction: inbound, outbound, dialer, internal

• channel: voice, chat, sms, WhatsApp, Openchannel, Cally Square

• queues

• dispositions

• Agents name (or IVR project name if it is a Cally Square)

• date range

Available widgets:

• Overall sentiment of the conversations analysed within the system (based on the filters entered). It shows the overall prevailing sentiment

• Sentiment Percentages display a breakdown of Sentiment, categorized into Positive, Negative, Neutral, and Mixed

• Sentiment Over Time: shows the Sentiment variation over time. By default, each square represents one day. Click on the graph to zoom in on the display.

• Agent Sentiment Widgets display the Sentiment details for each agent. It is also possible to order agents by a specific sentiment.

• Queue Sentiment displays the Sentiment details for each queue

Quality Analysis - Post Call Analytics

On this dashboard, with different widgets and filters, you can view provided data by the Post-Call Analytics feature from AWS and is available exclusively for the Voice channel (voice calls – inbound, outbound and dialer – and Cally Square).

Open

• Average Sentiment trend for both agents and customers. This widget considers the overall sentiment score visible on Motion in post-call analysis (it represents the average of the scores for all call quarters for all calls considered)

• Categories Compliance: the graph shows the AWS categories matched with their respective percentages compared to the number of calls analyzed. By creating categories, you can also use the category filter.

• Sentiment Trend Over Time for both agents and customers. By default, each square represents one day. Click on the graph to zoom in on the display.