5. Quality Assurance - Business Questions
How do I ensure compliance?
Increase Process Efficiency?
Increase customer service
and sales?
Reduce Headcount?
6. Quality Assurance - Performance Questions
Who needs help
and in what areas?
Is it delivering the
right levels of quality?
How are individual
staff performing?
Are they improving
over time?
How is my Contact
Centre performing?
7. • Who needs assistance?
• In what areas do they need
assistance?
Performance Questions
9. Record Call Transcribe Call Score Call
Commercial call
recording software
Transcription Engine Scoring Model
Report
Reporting and
Analytics Data Mart
How does it work ?
14. STRENGTHS
Spotting words and phrases in context2
Understanding workflow3
Fast to implement4
Sentiment & Charged Words5
Subjective Measures1
Soft Skills2
WEAKNESSES
NLP has advanced significantly1
Speed, Pace, Amplitude6
Multi-Vernacular calls3
15. “…Can I speak to
your manager…”
“…You not listening
to me…”
“…I am unhappy…”
“…I would like to
escalate this…”
“…Ask your
manager to call
me…”
“…I will report
you…”
“…I will sue you…”
“…I will contact the
ombudsman…”
“…you are harassing
me…”
Sentiment
17. Reports And analytics Slice and Dice Performance by
Campaign, By Agent and by Time
Identify Top and Poorly
performing Agents
Compare Call Centre
Quality performance by
Time- Drill Down to
Month, Week, Day
18. Reports And analytics View individual Agent Performance
over time
Obtain Summary
performance by Agent
by Section, By Time
19. Reports And analytics
Drill down into specific
calls to see breakdown
of average score by
section
Drill into each call
transcription for
proactive review and
training
21. Get an Engine that handles South African Accents an has a plan for Vernacular2
Make sure you can train your engine to increase accuracy3
Have a team good at AI - NLP4
60 % QA Cost savings
Increase QA samples by 10 - 50 X
5
Setup your recorder for dual channel1
Using ASR for QA - Summary
25. RPA and Virtual assistants to become the norm2
Contact centres in a position to build unique data asset5
Contact centres will decrease in size1
AI + Humans = Better Customer Service
4 People still want to be able to talk to a person if they are not being satisfied
3 Chatbot experience is key
28. Using AI / ML for Agent Customer Matching for better service2
Recommender engines for cross sell and up sell4
Combine Chatbots with service agents for faster service1
ML to extend contact centre support business
3 Prioritise calls and conversations for better conversion
29.
30. I’d love to chat with you
Jeff.Kian@eoh.co.za
JKian@xds.co.za
083-440-8129
Notes de l'éditeur
Jeff Kian
Head up a team for Big Data and Advanced Analytics
Build models for
Credit Risk Scoring
Marketing Services - Recommender Engines
Natural Language Processing
Big Data – ML for identifying and human trafficking
Supervised Learning
Human trains algorithm on labeled data to learn a relationship between an input & the output variable
Unsupervised Learning
Algorithm explores data to discover hidden patterns, groups and outliers
Reinforcement learning
Algorithm uses trial-and-error identify ideal behavior to maximize performance
Neural Networks
Group of algorithms structured in a network, to mimic the functions and structure of the brain