Skip to content

Speech Analytics: Insights From Your Customer’s Voice

Share on twitter
Share on facebook
Share on linkedin

Most enterprises experience the “moments of truth” at the time of customer service (for instance, a lost credit card, a canceled flight, a damaged piece of clothing, or investment advice) when customers invest a high amount of emotional energy in the outcome.  These interactions have a dramatic impact on the Customer experience. Analyzing these interactions can give deeper insights into process, performance, and customer experience issues.

Customer service centers are dominated by voice interactions between customers and service center agents, who are the face of the company.

The audio call data is archived for reference and not mined for customer interactions and extract value from them. It is exhausting to hear the tapes; hence this source of rich data is routinely ignored. Often this data is never accessed, unless there is a special situation such as an escalation or a dispute. Most information is hidden and should be mustered from the customer call data.  Since, it is Audio, many Enterprises give it the least preference.

However, competitive advantage of a company often depends on anticipating and addressing market needs faster than the competitors. Calls received by the service center can give a telling indicator about your business and market.

Speech Analytics should help organizations enrich customer interactions, improve business processes, and optimize their work forces to enhance loyalty, increase revenue, mitigate risk, and manage operational costs.

Speech Analytics can be of two variants:

  1. Synchronous Analysis: This is accomplished real-time and continuously monitored and tracked; which allows the service center the ability to provide superior customer engagement, handle situations at the right time to mitigate attrition and escalations.
  1. Archival Analysis: This is accomplished by mining intelligence from thousands of recorded calls/chats essential for pinpointing cost drivers, trends, and opportunities; identifying strengths and weaknesses with processes and products; and understanding how your offerings are perceived by the marketplace.

Core functions expected from a Speech Analytics are:

  • Real-time Mood Analysis – Analyze mood of the customer in a conversation (Emotions with a pre-defined time slices). Customer service will enhance tremendously by knowing the moods of the customer and intensity of conversation which will allows taking necessary actions to provide better service for the customer.
  • Agent Quality Scoring – Score service center agent on various compliance parameters to enhance the quality assurance practices. This will also help to identify training and resource re-allocation requirements.
  • Voice Transcription – Convert voice to text using Statistical machine translation techniques and Sentiment of the customer (Meaning)
  • Topic Detection: Provide sophisticated conversational analytics to automatically identify, group, and organize the words and phrases spoken during calls into themes, helping to reveal rising trends and areas of opportunity or concern. The solution should identify themes automatically, continually refine them, and add new themes over time.
  • Performance Analysis of the Service center: Should process thousands of unstructured data rapidly. Analysis of this information should provide information about customer sentiment, up-sell and cross-sell opportunities.

Using RapidMiner Studio as Speech Analytics Solution

The Cappius data science solution built with RapidMiner Studio processes customer communication channels like voice (calls) and text (chats) to decipher sentiments and mood of the conversation in a customer service call.

The analysis is processed in real time through continuous tracking, which gives service centers the ability to deliver superior customer engagement and handle critical situations at the right time to mitigate attrition and escalations. Our Audio processing supports conversational analytics and can separate interactions into agent and customer streams.

Cappius uses RapidMiner Studio for core data analysis, sentiment analysis, topic mining, agent scoring and as a predictive engine for trend analysis.

Benefits of the solution

  • Serves as an early warning system to identify issues before they escalate and impact a large number of customers
  • Helps you optimize customer engagement and service strategies by revealing trends and opportunities
  • Drives compliance by quickly revealing regulatory breaches and failure to adhere to internal policies

Speech Analytics has a great potential to enhance customer perception by pre-emptive actions that will result in preventing customer dissatisfaction and increase customer delight; achieving a superior service center performance.

Surya Putchala

Surya Putchala

Lead Big Data Analytics, Cappius Technologies Inc