© 2018 TM Forum | 1
Vice President, AI, Data Analytics & Customer Experience
Aaron Richard Earl Boasman-Patel
TM Forum AI Program: An Overview
© 2018 TM Forum | 2
1805
Thomas Bayes’ essay that
underpins AI
2011
IBM’s Watson beats Jeopardy!
2018
AI is a Reality!
Market Dynamics
AI: From Theory to Hype to Reality
Gartner Press Release 25 April, 2018
Business Value of AI
AI has great potential
to address core
challenges of the
telecom industry :
TM Forum Survey
Report
© 2018 TM Forum | 3
Artificial Intelligence could spur global growth as much
as the steam engine did…
• Artificial intelligence could contribute an additional 1.2 percent to annual GDP growth for
at least the next decade
• 70 percent of companies will adopt at least one form of AI by 2030, and a significant portion
of large firms will use a full range of the technology.
• Countries that establish themselves as AI leaders — mostly developed economies—
could capture 20 to 25 percent more in economic benefits compared to current levels*.
* Source: McKinsey Global Institute.
AI will add an extra $13trn
(£10trn) to global economic
activity by 2030, equivalent
to about 1.2% of additional
GDP growth a year
© 2018 TM Forum | 4
• 90% of the world’s data was generated
in the past two years
• Analysts forecast that by 2025 data will
exponentially grow by 10 times and
reach 163 zettabytes (Trillion gigabytes)
(Source IDC)
• Still a long way to go to harness the
power of this data. Approximately 1% of
data generated is utilized, processed
and acted upon
• Imagine what could happen if we were
able to effectively leverage and manage
more of this data at scale
AI is a $200bn opportunity to the data centre industry
© 2018 TM Forum | 5
CSPs need to embrace AI NOW!
AI is not new in the telecom industry. AT&T, for example, says it has been using the technology for
decades in areas like call center automation. But now it is imperative that CSPs adopt AI for two
primary reasons, both of which will lead to huge cost savings:
Modernizing management and operations
Without network automation the telecom business
model is at risk of breaking down. For networks to
support the billions of devices that are expected to be
connected to the internet within the next decade, they
must be self-optimizing and self-healing.
This requires machine learning.
Improving customer centricity
Within the last three to five years, improving customer
centricity has become the single biggest strategic
priority for telecom operators. AI is needed to give
customers the kinds of digital experiences they are
demanding, and it can deliver these capabilities
through chatbots and voice assistants.
© 2018 TM Forum | 6
AI Use Cases for Customer CentricityChatbotstocommunicatewithcustomers
Intelligent care use cases
Billing
Proactively
address billing-
related issues
such as an
abnormal fee, first
bill experience,
etc…
Devices
Support
self-installation of
new equipment,
configuration,
identify issues
with equipment,
predict
equipment
failures
Retention
Real-time churn
prevention –
identify behaviors
and patterns that
lead to potential
churn
Collection
Assist customers
in avoiding
negative billing
and payment
situations
Network
Real-time
detection of
quality-of-service
incidents, such as
network outages,
dropped calls,
etc…
* Source: TM Forum based on Amdocs illustration.
© 2018 TM Forum | 7
AI Use Cases for Customer CentricityChatbotstocommunicatewithcustomers
Intelligent marketing use cases
Onboarding
* Source: TM Forum based on Amdocs illustration.
Education &
guidance
Win back
Usage &
activity
simulation
Mobile app
engagement
Household
engagement
Price plan
migration
Digital service
adoption
Multi-play
cross-sell
Surprise &
delight
Loyalty
Handset
upgrade
Content and
OTT upsell
Household shared
services
Recommitment
Roaming
Onboarding Ongoing Churn prevention & retention Win back Cross stages
© 2018 TM Forum | 8
AI Use Cases for Network and Service Management
Autonomous networks have potential
* Source: TM Forum based on data from the World Economic Forum.
30%
Reduction in mobile
infrastructure spending
$46 billion
Savings in customer acquisition
costs and lost revenue through
better network performance
$9 billion
In operating profit from
reduced frequency and duration
of network outages
$14 billion
Generated from the sale of
self-organizing network solutions
30,000 tonnes
In CO2 emissions saved globally
because of fewer field visits
$12 billion
In consumer benefits from CSPs
passing network cost-savings on
$27 billion
Cumulative cost-savings for
the telecom industry over the
coming decade
2.5 billion hours
Saved and $3.8 billion in productivity
gains through reduction in dropped calls
© 2018 TM Forum | 9
What are the biggest drivers of AI for CSPs?
37%
15%
48%
What are the biggest drivers of AI?
Reducing OpExin the organisation by driving automation and closed loop systems
Upselling new products to existing customers
Delivering a better customer experience
* Source: TM Forum 2017.
© 2018 TM Forum | 10
AI Augmentation is the future
A study from Oracle and Future Workplace
shows that 93% of people would trust orders from a robot at work, while all the respondents agreed that AI will have a positive impact on their organization.
93% 59% 50%
Operational
efficiencies
Enabling faster
decision-making
Significantly
reduced costs
45% 37%
Enabling better
customer experiences
* Source: Oracle and Future Workplace.
Trust orders from
a robot
© 2018 TM Forum | 11
Huge Challenges to implementing and operationalising AI
Ethics
What are the biggest challenges?
1 2 3 4 5 6 7
CSPs
Suppliers
Lack of maturity of
network components
and support systems
Lack of software
expertise within
the organiztion
Overcoming fear
that automation will
limit control and
result in outages
Lack of data
analytics
expertise
within the
organization
Lack of
standards
for end-to-end
management
Concerns
about
security
Concerns
about
displacing
staff
Lack of maturity of
network components
and support systems
Lack of data
analytics
expertise
within the
organization
Lack of software
expertise within
the organization
Overcoming fear
that automation
will limit control
and result
in outages
Lack of
standards
for end-to-end
management
Concerns
about
security
Concerns
about
displacing
staff
* Source: TM Forum 2017.
© 2017 TM Forum | 12© 2018 TM Forum | 12
So how is TM Forum serving
the industry?
© 2018 TM Forum | 13
Challenge 1
Challenge 3
Challenge 2
Challenge 4
L A C K O F A I T R A I N I N G
R E P O S I T O R Y
Large pools of data are needed to train AI/ML
models. Everyone needs to contribute in a secure
fashion.
T H E R E A R E N O S T A N D A R D S
Managing AI at scale presents challenges of
accountability, audit and maintenance.
C R O S S V E R T I C A L A I D A T A
M O D E L S M I S S I N G
Need an Operational Telco Architecture open
enough to be able to work with cross-industry
standard AI/ML capabilities
W H E R E T O B E G I N , W H E R E T O
E N D
Difficult to articulate a vision for the journey of
implementing AI in CSP networks and operations.
Industry challenges in realizing the potential of AI
These challenges can only be overcome jointly
© 2018 TM Forum | 14
D A T A M O D E L
Requirements on the Open-
Digital Architecture to
support AI/ML
1
M A N A G E M E N T
S T A N D A R D S
Standard interfaces and
standard component
structures
2
T R A I N I N G
R E P O S I T O R Y
Data requirements, data
sourcing and data use
3
M A T U R I T Y
M O D E L
AI Maturity Model and
Methodology for CSPs to
assess AI adoption status
4
Addressing Member’s challenges through TM Forum AI Program: 5 Workstreams
5
A I U S E R S T O R I E S ( G U I D E B O O K )
• Interconnects the other collaboration projects
• It will collaboratively create and design small detailed and specific AI user stories and use cases
• Will include story maps, use cases, workflow diagrams, storyboards, sketches and mock-ups
5
© 2018 TM Forum | 15
View videos: https://tinyurl.com/ydy3249y
AI-related Catalyst Projects are proof of what can be achieved
© 2018 TM Forum | 16
Participating companies to date…
© 2018 TM Forum | 17
Key industry outcomes and benefit objectives
1 2 3 4 5
Create an
industry-agreed
and standardized
data model for AI
which will enable AI/ML
vendors to build
capabilities and
interoperable solutions
which will work across the
industry, and allow CSPs to
deliver enterprise
grade-services across
the digital ecosystem
enabled by AI.
Establish a set of
standard interfaces
for operational AI
Models
to allow them to be
understood monitored in
real-time using defined
architectural components
to develop and manage AI
models such as feature
databases, model
databases, deployment
servers and registration
databases. It will allow
service providers to bring
accountability to their AI
applications and allow
them to accurately audit,
maintain and withdraw
corrupted data sets quickly
and at low cost.
Enable a AI Data
Training Repository as
part of TM Forum’s
Open Labs project
This will allow members
to test their AI
applications in a safe
environment using large
sets of curated,
anonymized training data,
and make it easier to
test out the benefits
of different
vendor solutions.
Create an industry
standard AI
maturity model
and methodology
so that service providers
can assess how far along
the journey they are in
implementing AI in their
company, networks and
operations. It will be
aligned to TM Forum’s
Digital Maturity Model so
CSPs can consider AI
maturity in the context of
their overall digital
transformation assessment
and strategy.
Design and create an
industry-agreed set of
detailed and specific
AI user stories and
use cases
which will help the
entire industry focus
on the same priorities
when it comes to AI & ML.
© 2018 TM Forum | 18
The challenge is here…
The challenge is real, it is here and it is now!
© 2018 TM Forum | 19
Aboasman-patel@tmforum.org
Aaron Boasman-Patel
VP, AI, Data Analytics & Customer Centricity
Program
Join the AI Program
Go to:
https://www.tmforum.org/ai-data-analytics/#form

TM Forum AI Program Overview

  • 1.
    © 2018 TMForum | 1 Vice President, AI, Data Analytics & Customer Experience Aaron Richard Earl Boasman-Patel TM Forum AI Program: An Overview
  • 2.
    © 2018 TMForum | 2 1805 Thomas Bayes’ essay that underpins AI 2011 IBM’s Watson beats Jeopardy! 2018 AI is a Reality! Market Dynamics AI: From Theory to Hype to Reality Gartner Press Release 25 April, 2018 Business Value of AI AI has great potential to address core challenges of the telecom industry : TM Forum Survey Report
  • 3.
    © 2018 TMForum | 3 Artificial Intelligence could spur global growth as much as the steam engine did… • Artificial intelligence could contribute an additional 1.2 percent to annual GDP growth for at least the next decade • 70 percent of companies will adopt at least one form of AI by 2030, and a significant portion of large firms will use a full range of the technology. • Countries that establish themselves as AI leaders — mostly developed economies— could capture 20 to 25 percent more in economic benefits compared to current levels*. * Source: McKinsey Global Institute. AI will add an extra $13trn (£10trn) to global economic activity by 2030, equivalent to about 1.2% of additional GDP growth a year
  • 4.
    © 2018 TMForum | 4 • 90% of the world’s data was generated in the past two years • Analysts forecast that by 2025 data will exponentially grow by 10 times and reach 163 zettabytes (Trillion gigabytes) (Source IDC) • Still a long way to go to harness the power of this data. Approximately 1% of data generated is utilized, processed and acted upon • Imagine what could happen if we were able to effectively leverage and manage more of this data at scale AI is a $200bn opportunity to the data centre industry
  • 5.
    © 2018 TMForum | 5 CSPs need to embrace AI NOW! AI is not new in the telecom industry. AT&T, for example, says it has been using the technology for decades in areas like call center automation. But now it is imperative that CSPs adopt AI for two primary reasons, both of which will lead to huge cost savings: Modernizing management and operations Without network automation the telecom business model is at risk of breaking down. For networks to support the billions of devices that are expected to be connected to the internet within the next decade, they must be self-optimizing and self-healing. This requires machine learning. Improving customer centricity Within the last three to five years, improving customer centricity has become the single biggest strategic priority for telecom operators. AI is needed to give customers the kinds of digital experiences they are demanding, and it can deliver these capabilities through chatbots and voice assistants.
  • 6.
    © 2018 TMForum | 6 AI Use Cases for Customer CentricityChatbotstocommunicatewithcustomers Intelligent care use cases Billing Proactively address billing- related issues such as an abnormal fee, first bill experience, etc… Devices Support self-installation of new equipment, configuration, identify issues with equipment, predict equipment failures Retention Real-time churn prevention – identify behaviors and patterns that lead to potential churn Collection Assist customers in avoiding negative billing and payment situations Network Real-time detection of quality-of-service incidents, such as network outages, dropped calls, etc… * Source: TM Forum based on Amdocs illustration.
  • 7.
    © 2018 TMForum | 7 AI Use Cases for Customer CentricityChatbotstocommunicatewithcustomers Intelligent marketing use cases Onboarding * Source: TM Forum based on Amdocs illustration. Education & guidance Win back Usage & activity simulation Mobile app engagement Household engagement Price plan migration Digital service adoption Multi-play cross-sell Surprise & delight Loyalty Handset upgrade Content and OTT upsell Household shared services Recommitment Roaming Onboarding Ongoing Churn prevention & retention Win back Cross stages
  • 8.
    © 2018 TMForum | 8 AI Use Cases for Network and Service Management Autonomous networks have potential * Source: TM Forum based on data from the World Economic Forum. 30% Reduction in mobile infrastructure spending $46 billion Savings in customer acquisition costs and lost revenue through better network performance $9 billion In operating profit from reduced frequency and duration of network outages $14 billion Generated from the sale of self-organizing network solutions 30,000 tonnes In CO2 emissions saved globally because of fewer field visits $12 billion In consumer benefits from CSPs passing network cost-savings on $27 billion Cumulative cost-savings for the telecom industry over the coming decade 2.5 billion hours Saved and $3.8 billion in productivity gains through reduction in dropped calls
  • 9.
    © 2018 TMForum | 9 What are the biggest drivers of AI for CSPs? 37% 15% 48% What are the biggest drivers of AI? Reducing OpExin the organisation by driving automation and closed loop systems Upselling new products to existing customers Delivering a better customer experience * Source: TM Forum 2017.
  • 10.
    © 2018 TMForum | 10 AI Augmentation is the future A study from Oracle and Future Workplace shows that 93% of people would trust orders from a robot at work, while all the respondents agreed that AI will have a positive impact on their organization. 93% 59% 50% Operational efficiencies Enabling faster decision-making Significantly reduced costs 45% 37% Enabling better customer experiences * Source: Oracle and Future Workplace. Trust orders from a robot
  • 11.
    © 2018 TMForum | 11 Huge Challenges to implementing and operationalising AI Ethics What are the biggest challenges? 1 2 3 4 5 6 7 CSPs Suppliers Lack of maturity of network components and support systems Lack of software expertise within the organiztion Overcoming fear that automation will limit control and result in outages Lack of data analytics expertise within the organization Lack of standards for end-to-end management Concerns about security Concerns about displacing staff Lack of maturity of network components and support systems Lack of data analytics expertise within the organization Lack of software expertise within the organization Overcoming fear that automation will limit control and result in outages Lack of standards for end-to-end management Concerns about security Concerns about displacing staff * Source: TM Forum 2017.
  • 12.
    © 2017 TMForum | 12© 2018 TM Forum | 12 So how is TM Forum serving the industry?
  • 13.
    © 2018 TMForum | 13 Challenge 1 Challenge 3 Challenge 2 Challenge 4 L A C K O F A I T R A I N I N G R E P O S I T O R Y Large pools of data are needed to train AI/ML models. Everyone needs to contribute in a secure fashion. T H E R E A R E N O S T A N D A R D S Managing AI at scale presents challenges of accountability, audit and maintenance. C R O S S V E R T I C A L A I D A T A M O D E L S M I S S I N G Need an Operational Telco Architecture open enough to be able to work with cross-industry standard AI/ML capabilities W H E R E T O B E G I N , W H E R E T O E N D Difficult to articulate a vision for the journey of implementing AI in CSP networks and operations. Industry challenges in realizing the potential of AI These challenges can only be overcome jointly
  • 14.
    © 2018 TMForum | 14 D A T A M O D E L Requirements on the Open- Digital Architecture to support AI/ML 1 M A N A G E M E N T S T A N D A R D S Standard interfaces and standard component structures 2 T R A I N I N G R E P O S I T O R Y Data requirements, data sourcing and data use 3 M A T U R I T Y M O D E L AI Maturity Model and Methodology for CSPs to assess AI adoption status 4 Addressing Member’s challenges through TM Forum AI Program: 5 Workstreams 5 A I U S E R S T O R I E S ( G U I D E B O O K ) • Interconnects the other collaboration projects • It will collaboratively create and design small detailed and specific AI user stories and use cases • Will include story maps, use cases, workflow diagrams, storyboards, sketches and mock-ups 5
  • 15.
    © 2018 TMForum | 15 View videos: https://tinyurl.com/ydy3249y AI-related Catalyst Projects are proof of what can be achieved
  • 16.
    © 2018 TMForum | 16 Participating companies to date…
  • 17.
    © 2018 TMForum | 17 Key industry outcomes and benefit objectives 1 2 3 4 5 Create an industry-agreed and standardized data model for AI which will enable AI/ML vendors to build capabilities and interoperable solutions which will work across the industry, and allow CSPs to deliver enterprise grade-services across the digital ecosystem enabled by AI. Establish a set of standard interfaces for operational AI Models to allow them to be understood monitored in real-time using defined architectural components to develop and manage AI models such as feature databases, model databases, deployment servers and registration databases. It will allow service providers to bring accountability to their AI applications and allow them to accurately audit, maintain and withdraw corrupted data sets quickly and at low cost. Enable a AI Data Training Repository as part of TM Forum’s Open Labs project This will allow members to test their AI applications in a safe environment using large sets of curated, anonymized training data, and make it easier to test out the benefits of different vendor solutions. Create an industry standard AI maturity model and methodology so that service providers can assess how far along the journey they are in implementing AI in their company, networks and operations. It will be aligned to TM Forum’s Digital Maturity Model so CSPs can consider AI maturity in the context of their overall digital transformation assessment and strategy. Design and create an industry-agreed set of detailed and specific AI user stories and use cases which will help the entire industry focus on the same priorities when it comes to AI & ML.
  • 18.
    © 2018 TMForum | 18 The challenge is here… The challenge is real, it is here and it is now!
  • 19.
    © 2018 TMForum | 19 Aboasman-patel@tmforum.org Aaron Boasman-Patel VP, AI, Data Analytics & Customer Centricity Program Join the AI Program Go to: https://www.tmforum.org/ai-data-analytics/#form