Big data has been a hot topic for a few years, and predictive analytics is an aspect of data analytics that purports to forecast the outcome of future events. It does this by examining vast amounts of data, to identify patterns to inform the likely future outcomes of similar events. The power of artificial intelligence or machine learning, partnered with low cost cloud computing means processing huge  amount of data has become cost efficient and accessible to many. Importantly AI can continually take the feedback of real world results to ensure the models accuracy is constantly improving. Essentially the more data the better!

So where does speech analytics fit in to this?

Going beyond simple word or phrase searching, predictive analytics when applied to speech, can deliver highly informative insights. Applying a big data approach allows calls, in vast volumes, to be examined for the underlying patterns and complex events that would suggest a particular outcome. Let’s consider a few practical examples of how useful this could be:

  • Improving sales results: From sales enquiries, rapidly identifying the “hot” leads, the callers who are most likely to buy, or finding cross and upsell opportunities.
  • Customers at risk: Spotting callers who are considering cancelling their existing services, gives the supplier an opportunity to address the customers concerns.
  • Agent insight: Identify when an agent has used inappropriate language that breaches regulatory guidelines, has deviated from a script or missed key questions.
  • Automated agent call scoring: Build scorecards to check every aspect of every call, to replace human scoring of a handful of calls.
  • Customer insights: Discover if a caller mentions a competitor, asks about pricing, or raises a concern about support for example.
  • PCI detection & redaction: Whilst there are numerous solutions from simple manual  “stop/start” of recording to automated systems to manage PCI information, Vivo can provide a further layer of PCI compliance by checking calls for PCI data, and “scrubbing” any missed calls before returning them to the user.


For most call centres depending on human intervention to listen and assess calls , means results are reflective of a tiny sample of calls, and are inevitably subjective. Applying sophisticated predictive analytics techniques delivers results more quickly, more accurately & with greater insight.

Find out what you might be missing in your calls !

Andy Andreou

Sales Director