Speech Analytics has proven to be a powerful solution for optimizing workforce performance and the customer experience.

But, don’t just take my word for it! Look at some of the dramatic results that our customers have achieved with Speech Analytics, including:

  • 23% reduction in average handle time (AHT) while simultaneously improving customer satisfaction more than 11%
  • 18% improvement in first contact resolution (FCR)
  • 41% improvement in sales conversions

So, you may be wondering just how Speech Analytics can deliver these benefits. The secret lies in taking a broader view by analyzing all interactions, including text-based interactions. At Genesys, we call this Interaction Analytics, which delivers business results in three primary ways:

  1. Categorizing all conversations according to every topic discussed
  2. Automatically discovering emerging trends and unexpected events within the voice of the customer
  3. Enabling users to search for particular words or phrases within interactions

Although search is often considered the primary usage of Interaction Analytics and is definitely a key capability, categorization is the driver behind most of the business benefits that our customers have experienced.  Also, since “you don’t know what you don’t know,” it’s often crucial for an Interaction Analytics solution to be able you to automatically discover emerging trends or unexpected events that users wouldn’t have known to search for otherwise.

Determining Which Approach is Best for You

Speech Analytics GenesysThere are three kinds of Speech Analytics engines available today:

  1. Phonetic
  2. Speech-to-Text
  3. Speech-to-Phrase

Each of the above types of engines have distinct strengths and weaknesses, and depending on your use case, any of them can work. As a result, many vendors offer hybrid approaches when it comes to interaction analytics.  A good place to start is looking at each of your specific business requirement and evaluating which engine is best suited for your needs.

For example, phonetic engines are useful when searching for rare occurrences within large volumes of audio. However, phonetic engines are not very reliable (accurate and complete) for categorizing calls and it is impossible to automatically discover emerging trends or unexpected events using this approach. For this, speech-to-text transcription is required because Text Analytics can analyze the transcriptions to uncover those trends and events. Transcriptions can also be indexed to enable rapid ad-hoc searches and be merged with other text-based interactions (i.e. chat, email, social media) to enable unified analysis of all interactions across all channels.

The recommended best practice for maximizing these transcription-related benefits is to transcribe 100% of your organization’s calls into text.  Although speech-to-text engines are usually more reliable than phonetic engines for categorizing calls, both approaches begin by converting speech (into phonemes or text respectively), and some information is inevitably lost in conversion. To overcome this, speech-to-phrase engines can directly recognize entire phrases within the call audio, so no data is lost in conversion.  For this reason, speech-to-phrase engines are the most reliable for speech categorization.

Make Your Interaction Analytics Actionable

Interaction Analytics are ineffective unless they are actionable – meaning the analytics engine triggers actions that address the current issues or capitalize on the opportunities that are uncovered.  Below are some best practices for making your analytics actionable all the time:

  1. Begin by correlating your Key Performance Indicators (KPIs) against the interaction topics and agent skills identified by reliable interaction categorization.
  2. Leverage the information discovered to automatically trigger workflows to improve performance, such as automatically delivering e-learning to an agent on the specific topic or skill which the interaction analytics have uncovered as an area where the agent is underperforming.
  3. Measure the results of every action taken and tune your workflows and processes as needed to maximize your results.

To learn more about how to maximize the benefits of Interaction Analytics, see our eBook, Cancel Out the Noise:  Three Capabilities Every Speech & Text Analytics Solution Should Offer. You can read it here!

Sean Murphy

Sean Murphy

Sean Murphy has over 13 years of experience in the Analytics domain. Sean currently leads Product Marketing for Interaction Analytics at Genesys. Sean led Marketing at UTOPY, the Speech Analytics pioneer, for 4 ½ years before UTOPY was acquired by...