Skip to main content
Implementing Speech Analytics

Implementing Speech Analytics

We answer some questions posed by potential end-users of speech analytics technology as identified and highlighted in the Contact Babel 2014 Inner Guide to Customer Contact Analytics.

WHAT SHOULD WE CONSIDER WHEN CHOOSING A DEPLOYMENT MODEL FOR SPEECH ANALYTICS?
IS THE SAME FUNCTIONALITY AVAILABLE FOR CLOUD-BASED SOLUTIONS AND CPE?

The functionality for cloud and CPE based analytics technology is essentially the same. The deployment model you select can vary depending on your intended use or business model. For example outsourcers or seasonal based organisations may benefit from a hosted deployment if they only want to use the technology for a limited or specific period of time or for a specific client or project. A new trend we are also starting to see come through is speech analytics on a ‘managed service’ type basis. This is where an organisation may only seek use of speech analytics technology for a one off project or engagement to either get to the bottom of a specific issue or investigation. The technology is not owned by that organisation and a third party like Business Systems will take extracted call recordings and run them through an analytics engine and carry out the queries and analysis on that data and report back to the organisation on key findings and recommendations.

IS IT DIFFICULT TO IMPLEMENT SPEECH ANALYTICS IF A RECORDING VENDOR IS ALREADY PRESENT?
IF SO, HOW DO WE HANDLE THIS?

Yes it is true that interaction analytics can be more challenging to implement if a recording vendor is already present but there are always work-arounds and at Business Systems we have come up against this numerous times. As an independent supplier of both call recording and speech analytics solutions, we have worked with a variety of weird and wonderful combinations and have used tools like ‘extraction engines’ to take data out of an incumbent call recording solution and run it through whichever speech analytics system the client has opted for.

WHAT ARE THE MOST COMMON PITFALLS TO AVOID IN SPEECH ANALYTICS IMPLEMENTATIONS AND WHEN IN FULL PRODUCTION?

Not having objectives defined and set from the start can cause issues. Similarly not having the right resource in place, it’s not just about having people who can run the queries and generate the reports, they need to be able to draw conclusions and recommendations and establish processes and priorities for change management and see them through. Once initial goals and objectives have been reached the organisations that are prepared to change them and push for more will ensure that their ROI doesn’t plateau and that the speech analytics tool still remains dynamic and relevant. Finally speech analytics requires energy and drive and continual re-tuning to ensure it meets the changing dynamics of the environment in which it operates, it is not a ‘set-it-and-forget-it’ type technology which can be plugged in and left to run without any inputs.