Company Internal
1
Customer Happiness Score
Data Science & AI
Yettel Serbia:
Revolutionizing Telco Customer
Experience with AI
November 2025
Company Internal
Customer Care Challenges
AI as a Strategic
Enabler
o Reduces operational cost
o Improves customer satisfaction
(NPS/CSAT)
o Shortens resolution times
o Creates personalized experiences
o Unlocks proactive & predictive care
Complex Service Portfolio
(Mobile, Broadband, TV,
Bank..)
Customers expect a single point of contact
regardless of service. Agents must understand
many technologies (fiber optics, RF, DOCSIS, Wi-
Fi, VoIP).
Slow Insights Extraction
Time consuming insight generation from BSS (Billing, CRM,
Customer data) and OSS (Network configuration, provisioning,
quality metrics). Often 10–30 separate systems from different
vendors
Growing Expectation
Customers now expect telco support to
work like modern digial native apps .
Younger customers rarely want to call, and
prefer messaging apps, In-app support or
automated troubleshooting wizards
High call center volumes
Many issues are repetitive: SIM
activation, APN settings, router reboot,
billing questions. Network outages
create sudden spikes
Company Internal
Customer Care - Quality management and customer
satisfaction based on the sampling methodology
Next Best Action
Family of models predicting which care
activity will improve customers’
experience
Customer Happiness
Score
More advanced than CSAT or NPS –
estimates general customer happiness
and provides how customer experience
could be improved
Y.Buddy
Chatbots – both internal and external
leping customers and our agents to
optain complex information
Sonar
Internally developed ASR and text
processing tool helping to extract
insights from interactions with customer
across multiple touchpoints
Company Internal
SONAR – AI Automatic Speech Recognition
Audio
Two different CC
platform orovide two
different call recording
types with different
sampling rate and
channel quality
ASR
Internal data labeling.
Call augmentation
plays crucial roll.
Performance
improvement.
WER better compared
to some payed and
free off-shelf
solutions.
Text
Processing
Data labeling
performed with LLM,
and classification with
SLM – erformance
improvements.
Visualization
Different dashboards
must provide detailed
insights.
Anomaly detection will
help focus on
important changes.
E2E platform for ASR, text processing
and visualiyation
Company Internal
Y.Buddy – Chatbots
Customer Facing
o Multi-agent chatbots
o Built with both internally developed
LLM/SLM and vendor provided
models
o Help customer troubleshoot issues
and explore complex telco portfolio
o Provide dynamic personalized
information
o Perform provisioning
o Easy scalable chatbot platform
Internal
o Help agents navigate through
detailed internal procedures
o Lower time required to provide the
answer to customer
o Analyze information from multiple
systems
o Provide next best action
Company Internal
Customer Happiness Score
Profiling
Different customer profiles will be
influenced with differently by the same
events
BSS
Analyzing billing, tariff plans, voice and
data consumption estimate customer
satisfaction
Customer interactions
Analyze all written and verbal
interactions with customer evaluate
topics and sentiment
Service
Provide score estimation per customer
based on service quality
Company Internal
Next Best Action
AI Models
Churn model,
Customer happiness,
Sonar and many
other models provide
insights
Real-time
All systems must be
capable to perform
inference and
provide insights in
real-time
Presentation
AI guided support
must be presented in
structured and easy
to read form
Action
Process must suggest optimal
care or troubleshoot actions
across different channels. Every
actions must be recorded and
used as feedback loop for further
model improvments
Company Internal
8
Thank you!
Djordje Hirs

[DSC Europe 25] Djordje Hirs - Revolutionizing Telco Customer Experience with AI.pptx

  • 1.
    Company Internal 1 Customer HappinessScore Data Science & AI Yettel Serbia: Revolutionizing Telco Customer Experience with AI November 2025
  • 2.
    Company Internal Customer CareChallenges AI as a Strategic Enabler o Reduces operational cost o Improves customer satisfaction (NPS/CSAT) o Shortens resolution times o Creates personalized experiences o Unlocks proactive & predictive care Complex Service Portfolio (Mobile, Broadband, TV, Bank..) Customers expect a single point of contact regardless of service. Agents must understand many technologies (fiber optics, RF, DOCSIS, Wi- Fi, VoIP). Slow Insights Extraction Time consuming insight generation from BSS (Billing, CRM, Customer data) and OSS (Network configuration, provisioning, quality metrics). Often 10–30 separate systems from different vendors Growing Expectation Customers now expect telco support to work like modern digial native apps . Younger customers rarely want to call, and prefer messaging apps, In-app support or automated troubleshooting wizards High call center volumes Many issues are repetitive: SIM activation, APN settings, router reboot, billing questions. Network outages create sudden spikes
  • 3.
    Company Internal Customer Care- Quality management and customer satisfaction based on the sampling methodology Next Best Action Family of models predicting which care activity will improve customers’ experience Customer Happiness Score More advanced than CSAT or NPS – estimates general customer happiness and provides how customer experience could be improved Y.Buddy Chatbots – both internal and external leping customers and our agents to optain complex information Sonar Internally developed ASR and text processing tool helping to extract insights from interactions with customer across multiple touchpoints
  • 4.
    Company Internal SONAR –AI Automatic Speech Recognition Audio Two different CC platform orovide two different call recording types with different sampling rate and channel quality ASR Internal data labeling. Call augmentation plays crucial roll. Performance improvement. WER better compared to some payed and free off-shelf solutions. Text Processing Data labeling performed with LLM, and classification with SLM – erformance improvements. Visualization Different dashboards must provide detailed insights. Anomaly detection will help focus on important changes. E2E platform for ASR, text processing and visualiyation
  • 5.
    Company Internal Y.Buddy –Chatbots Customer Facing o Multi-agent chatbots o Built with both internally developed LLM/SLM and vendor provided models o Help customer troubleshoot issues and explore complex telco portfolio o Provide dynamic personalized information o Perform provisioning o Easy scalable chatbot platform Internal o Help agents navigate through detailed internal procedures o Lower time required to provide the answer to customer o Analyze information from multiple systems o Provide next best action
  • 6.
    Company Internal Customer HappinessScore Profiling Different customer profiles will be influenced with differently by the same events BSS Analyzing billing, tariff plans, voice and data consumption estimate customer satisfaction Customer interactions Analyze all written and verbal interactions with customer evaluate topics and sentiment Service Provide score estimation per customer based on service quality
  • 7.
    Company Internal Next BestAction AI Models Churn model, Customer happiness, Sonar and many other models provide insights Real-time All systems must be capable to perform inference and provide insights in real-time Presentation AI guided support must be presented in structured and easy to read form Action Process must suggest optimal care or troubleshoot actions across different channels. Every actions must be recorded and used as feedback loop for further model improvments
  • 8.