INTRODUCTION
Speech analytics is the process of transcribing an audio recording and analyzing the text for business intelligence use.
When an agent and customer are speaking, both sides of the conversation can be analyzed separately to determine various things such as:
MANUAL VS. AUTOMATION
Manual Review:
Subjective
Limited coverage (approx. 1% of calls)
Post-time analysis
Time-consuming
Speech Analytics:
Objective
Complete coverage (100% coverage)
Real-time analysis
Immediate feedback
WHAT IS POSSIBLE?
Call categorization
Evaluating agent performance
Determining customer experience
Assessing training effectiveness
Measuring risk
Addressing compliance violations
Detecting call avoidance
Discovering automation opportunities
Locating calls for calibration meetings
Identifying potential causes of:
High customer effort
Escalation calls
Customer confusion and frustration
LIMITATIONS
Speech analytics can:
Identify words that are said
Determine how those words are said
Determine when words are not said
Calculate silence time
Detect overtalk
Detect words that are in close proximity with the transcript
Enhance analysis with metadata such as:
Call direction
Wrap-up codes
Agent name
Speech analytics cannot:
Answer subjective questions
Transcribe some company names without going through an extra step
Determine which context is important
Determine if a response answers a question asked by a speaker
Determine if the agent took notes correctly