Four Ways Speech Transcription Can Transform Your Business

by Carmit DiAndrea
VP Portfolio Market Strategy
Verint
 
A few years ago, most conversations we had with speech analytics customers focused on features and functions of the technology.

These days, many of our speech analytics conversations focus on transcription accuracy. Features and functions still matter—particularly to contact centers—but new consumers of speech analytics data are emerging.  Big data and predictive analytics practitioners have enormous data sets, strong preferences for their tools of choice, and want access to the wealth of information that results from indexing and analyzing voice conversations. 

Here are some of the ways this rich data source can transform your business:

Using call transcripts to build NPS or CSAT models – For years, organizations have relied on post-call surveys to understand the customer experiences they deliver.  Surveys remain an invaluable source of direct customer feedback, but what if you could understand the customer experience on all calls, instead of only a subset of them?  And what if that understanding happened in such a timely fashion, you could take corrective action proactively to remedy less-than-optimal experiences? 

Big data and predictive teams are analyzing interactions with NPS or customer satisfaction (CSAT) survey results and building models to predict these scores for every interaction. This facilitates proactive outreach to customers whose experiences are under a defined threshold, offering a chance to improve CSAT and reduce customer churn (defection).

Using call transcripts to improve customer churn models – Churn is a complex and costly issue, and most large organizations have existing models that factor in key events triggering churn, churn behaviors of different customer demographics, etc. 

What’s missing from these models is the data from millions of conversations that take place between organizations and their customers. Speech analytics unlocks this data source via transcription (indexing of words) and enrichment (call categorization, identification of emotion, and understanding of conversational topics and word relations). Organizations can use this data to improve the predictive power of their models and decrease their rate of false positives.

Using call transcripts to build sales optimization models – Although every sales agent in a contact center might receive the same training, each develops his or her own unique style. And some are more effective than others.  Since even a one percent increase in sales conversion can be a game-changer, big data and predictive teams’ study and identify the most effective sales approaches, from word choice at key points of the conversation to conversational cadence and more.  Analyzing this data typically leads to a set of best practices and ineffective agent behaviors that are incorporated into sales training for new hires.
    
Using call transcripts as part of customer journey mapping initiatives – There are dozens of solutions available for bringing customer interactions and behavioral data together from different sources, allowing customer journeys to be traced across multiple engagement channels.

Call transcripts are key to understanding customer intent, successes, and challenges in the voice channel—as well as in digital channels, since customers typically default to calling when they have issues in other channels.  Why are so few customers using the newest self-service mechanism? The answer is likely contained in phone conversations. Why do customers who experience intermittent connectivity issues make an average of 2.5 calls on the subject? Same answer.

So, what’s next? Large-scale use of conversational data in predictive activities will most likely be followed by the use of prescriptive analytics to change the outcome of a conversation in real-time as the conversation is taking place.

Find out how Verint can help you leverage the latest advances in speech analytics and transcription to simplify, modernize, and automate customer engagement.