1. Helping people achieve a lifetime of financial security
Everyday VoiceAnalytics
Beata Steigerwald
Head Of Quality and Traffic Control
Aegon Hungary
25 September 2018
2. 2Aegon Contact Center
Contact Center environment
167 employees (154 FTE) working in 2 cities
71 Call Center agents; 61 Back Office workers (incl. 11 HomeOffice)
22 team managers & supporting collegues
800k calls + 970k client service traditional & e-documents on a yearly basis
3. 3
Our service challenge
Customer focused and cost effective in the same time!?
Aegon’s purpose
2017
Quality and performance
management and
churn reducing
Q3-Q4 2016 Quaility
management and
improving productivity
Apr 2016 Automatic
daily data refreshing
and daily alarms
Q1 2016
data upload
manually, and
analysis
Q4 2015.
3 month pilot
6. 6
- A combination of keyword and emotion
analysis
- Keywords like: our CEO, Hungarian
authorities, media
- 100% next-day control on the reported
conversations (on average 15-20 calls
per day 0,8% of ttl inbound calls)
- Immediate response, react to critical
calls, reducing escalating issues
Daily Alerts with customized reports
Everyday VoiceAnalytics
Reducing risk of authority and legal cases
7. 7Improving speech quality
Pilot for decreasing negative Emotion-Index
The positive, neutral, and negative categories defined by the program were used for
development. The initial values were high in the negative range.
• Our first goal was to reduce this and increase it to neutral values.
• As coaching cost EUR 150/agent resulted in suprisingly high success (avg 3,6 ->2,6)
8. 8
Control is on unproductive periods: 74% of void time caused by knowledge or
communication gap
o Time consuming to find or ask the right answer/solution
Resulted in decreasing unproductive periods
o Revealing individual training needs in product or/and systems, processes
o Improving communication skills
Rising efficiency and quality
Improving efficiency
Unproductive period is part of
the call center remuneration
system
HU financial sector avg. is 30,5%
Aegon CC avg was 30% at
launching the system
1% improvement saves 0,8 FTE
9. 9Everyday VoiceAnalytics
Increase productivity
• This system can filter out operators that have longer call time or void time above the average
• Efficiency and customer experience are enhanced by understanding what causes long void
time: ability or lack of knowledge may need to change the business process
• Only 2-3% of all cases can be controlled
11. 11
Emotions and tNPS
VoiceAnalytics feat. tNPS
Post integration analysis
shows that customer emotions
during calls are in line
with the surveys they fill out
after the call
We analyze and discover strong correlations
between conventional TNPS surveys and emotional
data readings.
12. 12
Call time vs tnps
Voiceanalytics feat tNPS
• Customers don’t appreciate long silence during calls, we have to be well prepared
for a call, and effective in our service
• The more complex a call, the less the customer is satisfied. Means: just keep it
simple and quick!
13. 13
Next steps
Using emotional data and tNPS correlation
analysis together with machine learning tools to
add tNPS value to every customer, based on a
conversation. This means tNPS score prediction
without any surveys!
Real time speech-analysis
Churn prediction and identification of upsell/xsell
opportunities during calls.
Next steps