VoiceAnalytics detects keywords such as invoice, contract, cancel, defect, return, problem, disaster, refund, lousy and expressions like for instance; ‘good morning,’‘ how can I help you?’, ‘please hold the line,’‘ thank you for your patience’ with high accuracy. It can also recognize industry-specific“out of directory” words like product names; it will recognize acronyms, coinage, events, campaigns or company names and other
names like those of agents, supervisors, and managers. These keywords can be entered, labeled and described phonetically through the user interface. Keyword spotting allows for listing and ranking conversations where these keywords were used, and it enables jumping to that exact point in the
conversation while playing back recordings.
Analytics and sentiment
The true emotion detection engine developed by Xdroid evaluates 30 speech markers to determine the emotions in a call. VoiceAnalytics analyses speech-blocks of a few seconds, and assigns an emotion value to it whenever an emotion, other than neutral, is detected. When this occurs, the system puts the emotion into one of four possible categories; happy, disappointed, displeased, uncertain, and assigns a value to it.
To achieve an overall emotion index of the conversation, a weighted sum over all Emotion Values of all speech-blocks is applied. The Emotion Index allows for a ranking of calls, from a high to low displeased emotion index. This enables comparing a call’s Emotion Index to the average value across all callsof
the contact center. The individual operator’s score can also easily be compared to the average rating of the group.
The VoiceAnalyticsand TextAnalyticsdashboards allows users to quickly drill down from the high-level overview individual voice files(which contains analysis data like emotions and keyword occurrences).
On top of the most commonly used KPIs for contact centers, like average response time and percentage of missed calls, custom KPIs and thresholds can be set-up to identify conversations with extreme values.
Another key advantage of VoiceAnalyticsis the combination of keyword spotting with true emotion analysis. It allows for a far more detailed and accurate qualification of the conversation than when using keyword spotting and keyword meaning only. Through the usage of our emotion analytics module, customers have found that the vast majority of “Thank you” expressions is an obligatory politeness statement rather than a sincere expression of gratitude and positive feelings. Half the time “Thank you” was expressed with negative emotions(or sarcasm)and evenly as many times “Thank you” was expressed with positive emotion (sincere thanks).
VoiceAnalytics supports the quality management of a contact center by quickly and easily identifying critical conversations based on its analytics data. It does so, for instance, by identifying dissatisfied customers, potential legal- and authority cases or by detecting customers that plan to cancel their contract.
It also tracks the individual scores of an agent or group of agents against the companies performance goals. VoiceAnalytics provides objective measurements that help identify the coaching and training needs so that appropriate action can be taken to improve communication and skills of an agent.
When combining analysis results with CRM and NPS data, VoiceAnalytics becomes a powerful quality management tool to support both customer and agent satisfaction retention.
CRM and TNPS
VoiceAnalytics supports assigning the value of CRM data and transactional net promoter score (TNPS) to each call, along with other information from, for instance, accompanied questionnaires. This can also be combined with emotions or other analytics data of the productivity of the conversation, in order to perform an in-depth analysis to identify drivers of TNPS, to improve the overall score.
Our users have experienced that there are more negative emotions and non-productive periods amongst the detractors. VoiceAnalytics leads to a higher focus on actions that can impact these KPIs, which in turn leads to a higher number of promoters.
The insight learning module, based on artificial intelligence, can perform analysis across all conversations to identify trends and patterns in the analytics data. This is especially important for conversations that ended in legal actions or for fraud attempt related conversations, etc.
Xdroid helps to recognize the factors that played a crucial role in the conversations that ended badly. It takes CRM data and speech characteristics into account, as well as complex relationships that, without MachineLearning, would be left uncovered.
The system will automatically generate rules that can be applied to new conversations, and alerts which can be set for conversations that meet these criteria, so that the right people can be notified in time and can take appropriate action.
With this approach, new conversations that show a particular trend can be identified far more efficiently. By combining multiple rules, MachineLearning can achieve even higher efficiency, so predicting the customer’s behavior is even more accurate.
At this moment, VoiceAnalytics delivers near real-time results. Depending on the hardware configuration, analytics data can already be given within minutes after uploading the call recordings, thereby providing valuable feedback to the contact center management in near real-time. The next generation of machine learning and artificial intelligence engines are implementing the most advanced deep neural networks to use machine learning algorithms that maximize and optimize GPU usage and performance. Those engines will support real-time analysis and give input to real-time agent coaching and bots.
Xdroid continuously improves and extends its VoiceAnalytics solution to meet and exceed customer expectations and has a number of great features and functionalities under development for commercial availability late 2019.
Advanced Scorecard / Scoreboard Capabilities
Many call centers use human labor to evaluate calls, simply by playing them back. VoiceAnalytics offers automated rules like the “greeting rule” which can check whether an agent properly greeted the customer or the “explore rule” which can check the number of questions an agent needed to ask to get the right information from the customer etc.
Xdroid is developing scorecard & scoreboard functionalities to automatically evaluate how each call scores against certain business rules and continuously adds support for more complex business rules.
In order for this functionality to work properly, both high sensitivity and high precision are needed. When measured against current external 3rd party engines, Xdroid’s next-generation engine provides very high hit rates.
Human resource module
This module is targeted at assisting HR management in criticalareas such as:
- Difficult hour – bad day
This functionality notifies managers when an agent had to handle several calls in a row with a negative Emotion Index, allowing the manager to intervene by giving an agent a well-needed break.
- Best fit in a group
When interviewing an agent, the system can show how much the speech characteristics of the potential new employee correlate with the best agents of several groups (sales agents, complaint handling agents, debt collection agents, etc.) so that a new hire is assigned to the group that fits best.
- Burn out alert
By collecting information about agents that evolved towards burnout and analyzing similarities in changes of the voice characteristics, burnout signals is detected early on.
- Training advisor
By comparing voice characteristics, keyword usage and productivity markers of an agent, with those of successful agents, training advice can be given for each agent or group of agents.