SlideShare una empresa de Scribd logo
1 de 30
Descargar para leer sin conexión
The Scope for Robotic Process Automation
& Machine Learning in Telecom Operations
James Crawshaw
Senior Analyst
Heavy Reading
First, the earth cooled …
Strowger switch (1891):
“Girl-less, cuss-less, out-of-order-less, wait-less telephone".
First exchange (1877):
Built from "carriage bolts, handles from teapot lids and bustle wire"
Evolution of automation
Time
Degree of automation
Electromechanical switch
Electronic switch
OSS
Telephone exchange
SDN, NFV, IBN, OMG
Brutal Automation
Ferocious, Savage, Bloodcurdling Automation
Intelligent process automation per McKinsey
1. RPA: software tool that automates routine tasks such as data
extraction and cleaning. Robot has user ID and performs rules-
based tasks such as accessing programs and systems, performing
calculations, creating reports, and checking files.
2. ML: algorithms that identify patterns in structured data, through
supervised or unsupervised learning, and provide insights or make
predictions.
3. NLP: a way for computers to analyze, understand, and derive
meaning from human language.
Robotics Process Automation
RPA use case
Source: Ovum
Typical RPA use cases
Use case Challenge Outcome
Customer Service &
Support Desk
Average time for 1st line support to
execute user admin tasks = 6 mins
Task time reduced to 50 seconds with
2 week project saving $1m annually
IT & Infrastructure
Support
Managed service provider
Needed a skilled support solution expert
for a 24/7 operation.
Bot created to handle admin tasks
including validation, regular testing,
diagnostics and fault remediation.
Data Migration &
Management
12 employees assigned to manually rekey
records from one system to the other.
2 bots created to replace 12 FTE
Back Office
Administration
300 staff using 18-year old mainframe
system and spreadsheets to process
queries in 20 minutes
Process time reduced to 2 minutes; 12
FTE saving; investment payback 2
months.
Connecting Process
Islands
HR starter/leaver process used disparate
systems with email handoff between
departments
Processing times reduced by 90%
Source: Thoughtonomy
Examples of RPA in telecoms
• Telefonica UK - SIM swaps, credit checks, order processing, customer
reassignment, unlatching, porting, ID generation, customer dispute
resolution and customer data updates.
• AT&T - migration of customer accounts from DirectTV, completing
requirement documents for Ethernet services, automation of sales
order entry, reconciliation of revenue against assets and inventory.
• Deutsche Telekom - 1,000 robots supporting 30m transactions per
year and saving 800 FTE
• Tele Danmark (TDC) - Customer facing (order handling and customer
care), Operations (updating drawings used in network planning),
Finance
• Elisa - automated pay-TV ordering; payback 3 months.
Specialized software versus BPMS vs RPA
Source: Blue Prism
RPA conclusions
• RPA is effective for processes that require predictable and high
frequency interactions with multiple applications.
• RPA is highly scalable - a robotic workforce can be doubled almost
instantly when new products are about to be launched, and then
scaled back after the surge.
• RPA can be a cost-effective alternative to Business Process
Outsourcing of back office functions.
• Telecom operations include many mundane and repetitive but
essential processes that require multiple systems to be queried
and/or updated to complete the task. The tasks must be completed
reliably and accurately - a textbook case for RPA.
Machine Learning
ML vs AI
“Machine learning is a subfield of AI, but it’s grown so large and
successful that it now eclipses its proud parent.” – Pedro Domingos,
The Master Algorithm
AI
ML
Deep
learning
E.g. pattern recognition, statistical modeling,
data mining, predictive analytics, data
science, self-organizing systems, etc.
E.g. Multilayer
Perceptron
E.g. Knowledge base
Machine Learning types
• Supervised learning: in addition to the data inputs we also show the
system the desired outputs and ask the system to create its own
mapping of inputs to output conditions.
• Classification systems e.g. Bayesian networks; Decision trees; Logistic
regression; Random Forest; Support Vector Machine.
• Regression systems e.g. KNN; Linear regression. A regression algorithm
indicates a statistical relationship between two, or more, variables (e.g.
temperature and noise).
• Unsupervised learning: system tries to find patterns in the input data
without knowing of any specific output conditions of the data.
Clustering techniques are normally applied to unsupervised learning
systems. Examples of clustering algorithms include Apriori;
Distribution-based (Gaussian mixture models); K-Means.
Why the resurgence of interest in ML?
• Breakthroughs in neural network theory around 2006
• Improvements in computing capacity: x86 CPUs, GPUs, FPGAs and
custom ASICs designed specifically for ML e.g. Google's Tensor
Processing Unit (TPU) and associated Tensor Flow software libraries.
• The cloud makes computing capacity highly available and cheap.
• Massive data sets: online photos, email, video, gaming, search,
messaging, mapping and shopping are fertile ground for ML.
• Success stories: AlphaGo, Google Pixel Buds, image recognition, lip
reading, etc.
Adoption of ML in telecoms
29%
31%
22%
18%
Are you deploying AI?
We're taking a wait and see approach
We have started proofs of concept
We are working with suppliers to build AI intro some products and services
We have built some internal AI expertise and are incoporating it into product and service roadmaps
Source: TMForum, 2017
Potential use cases for ML
• Customer facing
• chatbots
• contact center optimization and compliance
• sentiment analysis
• churn prediction and prevention
• fraud detection and prevention
• Networking
• anomaly detection for OAM&P (operations, administration, maintenance and
provisioning)
• security – e.g., threat detection
• automated resolution of trouble tickets
• prediction of network faults
• performance monitoring and optimization
• route optimization – e.g. in IP transport or SDN switching
• traffic identification – e.g. for policy enforcement
• SLA monitoring and enforcement
AI use cases in telecoms
Source: TMForum, 2017
ML and analytics for network management
Source: TMForum, 2017
ML in network management
• Vodafone trialed machine learning in a Centralized Self-
Organizing Network to identify the optimal settings to
deliver VoLTE and to predict traffic hotspots.
• KDDI developed an AI-based monitoring system to predict
anomalies caused by hardware and software.
• Zhejiang Telecom implemented AI engine for route
optimization, capacity planning, traffic prediction and
dynamic optimization of the network.
Barriers to ML in networking
• Network engineers typically don’t have backgrounds that
include the kinds of mathematical training and experience
that are essential in ML.
• Lack of good data sets: most data sets are noisy,
unnormalized. Most data sets are proprietary. There is no
standard way to integrate network data with other data
sources like CPU usage, memory, etc.
• No useful “Theory of Networking”
ML in DT’s RT-NSM Architecture
Billions of
events
Entropy
filtering,
deep learning
neural
networks
Filter
Event
categorization
Advanced
clustering:
K-Means, k-NN,
mixture models,
CURE, deep
learning, etc.
Recomm
endation
engine
Operations &
optimization
engine based
on deep
learning
algorithms
Reinforcement
learning
Continuous
improvement of
operational
performance
ML code is only a small part of the job
Source: Hidden Technical Debt in Machine Learning Systems, Sculley et al.
Knowledge-Defined Networking
Source: Knowledge-Defined Networking, Mestres et al
ETSI’s Experiential Networked Intelligence
Telecom Infra Project AI/ML working group
• Will focus on application of machine-based decisioning and auto-
remediation to help carriers keep pace with the growth in network
size, traffic volume and service complexity.
• Three work streams:
• ML-based network operations, optimization and planning to enhance
intelligence in network operations areas through, for example, predictive
maintenance and dynamic resource allocation
• Customer behavior-driven service optimization to enhance the overall
customer experience, particularly for bandwidth-intensive, latency-sensitive
and/or data-heavy applications
• Multi-Vendor ML-AI Data Exchange Formats to ensure the development of
generalized ML models that are applicable across the industry.
Automation
The Route To Automation
Traditional networking Cloud networking
Manual Software-driven
Delivery time 10-20 days (enterprise service) Delivery time 5-10 minutes
Static (>1 year contracts) Dynamic (on demand services)
Source: Colt
Automation end game
If we virtualize all
our networks
functions and use
big data analytics
and deep learning
algorithms to run
the operations …
… the entire
business can be
managed by just
one monkey …
… plus a second
monkey to look at
the PowerPoint
slides of the first
monkey.
Paradox of Automation
• “The more advanced a control system is, so the more crucial may be
the contribution of the human operator.” - Dr. Lisanne Bainbridge,
Department of Psychology, University College London.
• “We appear to be locked into a cycle in which automation begets the
erosion of skills or the lack of skills in the first place and this then
begets more automation” - William Langewiesche, ex professional
airplane pilot.
• “To err is human, to really foul things up requires a computer.” -
William E. Vaughan, newspaper columnist.
Q&A

Más contenido relacionado

Similar a The Scope for Robotic Process Automation & Machine Learning in Telecom Operations

Stop Wasting Energy on M2M
Stop Wasting Energy on M2MStop Wasting Energy on M2M
Stop Wasting Energy on M2MEurotech
 
IoT and the Oil & Gas industry at M2M Oil & Gas 2014 in London
IoT and the Oil & Gas industry at M2M Oil & Gas 2014 in LondonIoT and the Oil & Gas industry at M2M Oil & Gas 2014 in London
IoT and the Oil & Gas industry at M2M Oil & Gas 2014 in LondonEurotech
 
On the Application of AI for Failure Management: Problems, Solutions and Algo...
On the Application of AI for Failure Management: Problems, Solutions and Algo...On the Application of AI for Failure Management: Problems, Solutions and Algo...
On the Application of AI for Failure Management: Problems, Solutions and Algo...Jorge Cardoso
 
Rise of the machines -- Owasp israel -- June 2014 meetup
Rise of the machines -- Owasp israel -- June 2014 meetupRise of the machines -- Owasp israel -- June 2014 meetup
Rise of the machines -- Owasp israel -- June 2014 meetupShlomo Yona
 
From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...
From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...
From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...Rick G. Garibay
 
Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...
Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...
Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...Eurotech
 
BsidesLVPresso2016_JZeditsv6
BsidesLVPresso2016_JZeditsv6BsidesLVPresso2016_JZeditsv6
BsidesLVPresso2016_JZeditsv6Rod Soto
 
System Support for Internet of Things
System Support for Internet of ThingsSystem Support for Internet of Things
System Support for Internet of ThingsHarshitParkar6677
 
Design of an Autonomous Management and Orchestration for Fog Computing
Design of an Autonomous Management and Orchestration for Fog ComputingDesign of an Autonomous Management and Orchestration for Fog Computing
Design of an Autonomous Management and Orchestration for Fog ComputingSabelo Dlamini
 
Metering & Billing Audit
Metering  & Billing AuditMetering  & Billing Audit
Metering & Billing AuditDaniel Peter
 
Dell AI Telecom Webinar
Dell AI Telecom WebinarDell AI Telecom Webinar
Dell AI Telecom WebinarBill Wong
 
Distributed Computing system
Distributed Computing system Distributed Computing system
Distributed Computing system Sarvesh Meena
 
Applicability of Machine learning in computer networks
Applicability of Machine learning in computer networksApplicability of Machine learning in computer networks
Applicability of Machine learning in computer networksChetan Kumar S
 
Architecture & data acquisition by embedded systems in automobiles seminar ppt
Architecture & data acquisition by embedded systems in automobiles seminar pptArchitecture & data acquisition by embedded systems in automobiles seminar ppt
Architecture & data acquisition by embedded systems in automobiles seminar pptAnkit Kaul
 

Similar a The Scope for Robotic Process Automation & Machine Learning in Telecom Operations (20)

Stop Wasting Energy on M2M
Stop Wasting Energy on M2MStop Wasting Energy on M2M
Stop Wasting Energy on M2M
 
IoT and the Oil & Gas industry at M2M Oil & Gas 2014 in London
IoT and the Oil & Gas industry at M2M Oil & Gas 2014 in LondonIoT and the Oil & Gas industry at M2M Oil & Gas 2014 in London
IoT and the Oil & Gas industry at M2M Oil & Gas 2014 in London
 
Internet of things
Internet of thingsInternet of things
Internet of things
 
Internet of things
Internet of thingsInternet of things
Internet of things
 
On the Application of AI for Failure Management: Problems, Solutions and Algo...
On the Application of AI for Failure Management: Problems, Solutions and Algo...On the Application of AI for Failure Management: Problems, Solutions and Algo...
On the Application of AI for Failure Management: Problems, Solutions and Algo...
 
I learning lot
I learning lotI learning lot
I learning lot
 
Rise of the machines -- Owasp israel -- June 2014 meetup
Rise of the machines -- Owasp israel -- June 2014 meetupRise of the machines -- Owasp israel -- June 2014 meetup
Rise of the machines -- Owasp israel -- June 2014 meetup
 
From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...
From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...
From the Internet of Things to Intelligent Systems A Developer's Primer - Gar...
 
Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...
Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...
Addressing the Complexity and Risks of M2M Projects - M2M World Congress Apri...
 
BsidesLVPresso2016_JZeditsv6
BsidesLVPresso2016_JZeditsv6BsidesLVPresso2016_JZeditsv6
BsidesLVPresso2016_JZeditsv6
 
System Support for Internet of Things
System Support for Internet of ThingsSystem Support for Internet of Things
System Support for Internet of Things
 
Design of an Autonomous Management and Orchestration for Fog Computing
Design of an Autonomous Management and Orchestration for Fog ComputingDesign of an Autonomous Management and Orchestration for Fog Computing
Design of an Autonomous Management and Orchestration for Fog Computing
 
Metering & Billing Audit
Metering  & Billing AuditMetering  & Billing Audit
Metering & Billing Audit
 
Dell AI Telecom Webinar
Dell AI Telecom WebinarDell AI Telecom Webinar
Dell AI Telecom Webinar
 
Cloud computing
Cloud computingCloud computing
Cloud computing
 
Distributed Computing system
Distributed Computing system Distributed Computing system
Distributed Computing system
 
TpM2013: Bruno Albietz : A Hotel in the Cloud
TpM2013: Bruno Albietz : A Hotel in the CloudTpM2013: Bruno Albietz : A Hotel in the Cloud
TpM2013: Bruno Albietz : A Hotel in the Cloud
 
Applicability of Machine learning in computer networks
Applicability of Machine learning in computer networksApplicability of Machine learning in computer networks
Applicability of Machine learning in computer networks
 
Design of network
Design of networkDesign of network
Design of network
 
Architecture & data acquisition by embedded systems in automobiles seminar ppt
Architecture & data acquisition by embedded systems in automobiles seminar pptArchitecture & data acquisition by embedded systems in automobiles seminar ppt
Architecture & data acquisition by embedded systems in automobiles seminar ppt
 

Más de James Crawshaw

Open source for NFV MANO
Open source for NFV MANOOpen source for NFV MANO
Open source for NFV MANOJames Crawshaw
 
Advanced Analytics & AI for Network Operations
Advanced Analytics & AI for Network OperationsAdvanced Analytics & AI for Network Operations
Advanced Analytics & AI for Network OperationsJames Crawshaw
 
Software driven operations
Software driven operationsSoftware driven operations
Software driven operationsJames Crawshaw
 
Orchestrating the lifecycle of network services
Orchestrating the lifecycle of network servicesOrchestrating the lifecycle of network services
Orchestrating the lifecycle of network servicesJames Crawshaw
 
How CSPs are Leveraging Open APIs
How CSPs are Leveraging Open APIsHow CSPs are Leveraging Open APIs
How CSPs are Leveraging Open APIsJames Crawshaw
 
Software Defined Operations Research Presentation
Software Defined Operations Research PresentationSoftware Defined Operations Research Presentation
Software Defined Operations Research PresentationJames Crawshaw
 
Service assurance for NFV
Service assurance for NFVService assurance for NFV
Service assurance for NFVJames Crawshaw
 
Lessons learned so far in operationalizing NFV
Lessons learned so far in operationalizing NFVLessons learned so far in operationalizing NFV
Lessons learned so far in operationalizing NFVJames Crawshaw
 
Enabling the transition from CSP to DSP with B/OSS transformation
Enabling the transition from CSP to DSP with B/OSS transformationEnabling the transition from CSP to DSP with B/OSS transformation
Enabling the transition from CSP to DSP with B/OSS transformationJames Crawshaw
 
Industry progress towards a next gen oss for a virtualized network
Industry progress towards a next gen oss for a virtualized networkIndustry progress towards a next gen oss for a virtualized network
Industry progress towards a next gen oss for a virtualized networkJames Crawshaw
 
Industry progress towards a Next Gen OSS for a Virtualized Network
Industry progress towards a Next Gen OSS for a Virtualized NetworkIndustry progress towards a Next Gen OSS for a Virtualized Network
Industry progress towards a Next Gen OSS for a Virtualized NetworkJames Crawshaw
 
Is your digital transformation driving customer experience?
Is your digital transformation driving customer experience?Is your digital transformation driving customer experience?
Is your digital transformation driving customer experience?James Crawshaw
 
Automation, Agility and NFV
Automation, Agility and NFVAutomation, Agility and NFV
Automation, Agility and NFVJames Crawshaw
 
Closed Loop Automation for NFV
Closed Loop Automation for NFVClosed Loop Automation for NFV
Closed Loop Automation for NFVJames Crawshaw
 
Automation Opportunities in Telecom Operations With Intent-Based Networking, ...
Automation Opportunities in Telecom Operations With Intent-Based Networking, ...Automation Opportunities in Telecom Operations With Intent-Based Networking, ...
Automation Opportunities in Telecom Operations With Intent-Based Networking, ...James Crawshaw
 
Open source reality check
Open source reality checkOpen source reality check
Open source reality checkJames Crawshaw
 
Let's get the machines to do it
Let's get the machines to do itLet's get the machines to do it
Let's get the machines to do itJames Crawshaw
 
2020 2017 presentation analyst - james crawshaw
2020 2017 presentation   analyst - james crawshaw2020 2017 presentation   analyst - james crawshaw
2020 2017 presentation analyst - james crawshawJames Crawshaw
 

Más de James Crawshaw (19)

Open source for NFV MANO
Open source for NFV MANOOpen source for NFV MANO
Open source for NFV MANO
 
Advanced Analytics & AI for Network Operations
Advanced Analytics & AI for Network OperationsAdvanced Analytics & AI for Network Operations
Advanced Analytics & AI for Network Operations
 
Software driven operations
Software driven operationsSoftware driven operations
Software driven operations
 
Orchestrating the lifecycle of network services
Orchestrating the lifecycle of network servicesOrchestrating the lifecycle of network services
Orchestrating the lifecycle of network services
 
Zero Touch Analytics
Zero Touch AnalyticsZero Touch Analytics
Zero Touch Analytics
 
How CSPs are Leveraging Open APIs
How CSPs are Leveraging Open APIsHow CSPs are Leveraging Open APIs
How CSPs are Leveraging Open APIs
 
Software Defined Operations Research Presentation
Software Defined Operations Research PresentationSoftware Defined Operations Research Presentation
Software Defined Operations Research Presentation
 
Service assurance for NFV
Service assurance for NFVService assurance for NFV
Service assurance for NFV
 
Lessons learned so far in operationalizing NFV
Lessons learned so far in operationalizing NFVLessons learned so far in operationalizing NFV
Lessons learned so far in operationalizing NFV
 
Enabling the transition from CSP to DSP with B/OSS transformation
Enabling the transition from CSP to DSP with B/OSS transformationEnabling the transition from CSP to DSP with B/OSS transformation
Enabling the transition from CSP to DSP with B/OSS transformation
 
Industry progress towards a next gen oss for a virtualized network
Industry progress towards a next gen oss for a virtualized networkIndustry progress towards a next gen oss for a virtualized network
Industry progress towards a next gen oss for a virtualized network
 
Industry progress towards a Next Gen OSS for a Virtualized Network
Industry progress towards a Next Gen OSS for a Virtualized NetworkIndustry progress towards a Next Gen OSS for a Virtualized Network
Industry progress towards a Next Gen OSS for a Virtualized Network
 
Is your digital transformation driving customer experience?
Is your digital transformation driving customer experience?Is your digital transformation driving customer experience?
Is your digital transformation driving customer experience?
 
Automation, Agility and NFV
Automation, Agility and NFVAutomation, Agility and NFV
Automation, Agility and NFV
 
Closed Loop Automation for NFV
Closed Loop Automation for NFVClosed Loop Automation for NFV
Closed Loop Automation for NFV
 
Automation Opportunities in Telecom Operations With Intent-Based Networking, ...
Automation Opportunities in Telecom Operations With Intent-Based Networking, ...Automation Opportunities in Telecom Operations With Intent-Based Networking, ...
Automation Opportunities in Telecom Operations With Intent-Based Networking, ...
 
Open source reality check
Open source reality checkOpen source reality check
Open source reality check
 
Let's get the machines to do it
Let's get the machines to do itLet's get the machines to do it
Let's get the machines to do it
 
2020 2017 presentation analyst - james crawshaw
2020 2017 presentation   analyst - james crawshaw2020 2017 presentation   analyst - james crawshaw
2020 2017 presentation analyst - james crawshaw
 

Último

SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024Lorenzo Miniero
 
SAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptxSAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptxNavinnSomaal
 
DSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningDSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningLars Bell
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Commit University
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxhariprasad279825
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationSlibray Presentation
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Mark Simos
 
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxMerck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxLoriGlavin3
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsMiki Katsuragi
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLScyllaDB
 
Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Manik S Magar
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii SoldatenkoFwdays
 
From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .Alan Dix
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebUiPathCommunity
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek SchlawackFwdays
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 3652toLead Limited
 
The Ultimate Guide to Choosing WordPress Pros and Cons
The Ultimate Guide to Choosing WordPress Pros and ConsThe Ultimate Guide to Choosing WordPress Pros and Cons
The Ultimate Guide to Choosing WordPress Pros and ConsPixlogix Infotech
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brandgvaughan
 

Último (20)

SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024
 
SAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptxSAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptx
 
DSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningDSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine Tuning
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptx
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck Presentation
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
 
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxMerck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering Tips
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQL
 
Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko
 
From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio Web
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365
 
The Ultimate Guide to Choosing WordPress Pros and Cons
The Ultimate Guide to Choosing WordPress Pros and ConsThe Ultimate Guide to Choosing WordPress Pros and Cons
The Ultimate Guide to Choosing WordPress Pros and Cons
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brand
 

The Scope for Robotic Process Automation & Machine Learning in Telecom Operations

  • 1. The Scope for Robotic Process Automation & Machine Learning in Telecom Operations James Crawshaw Senior Analyst Heavy Reading
  • 2. First, the earth cooled … Strowger switch (1891): “Girl-less, cuss-less, out-of-order-less, wait-less telephone". First exchange (1877): Built from "carriage bolts, handles from teapot lids and bustle wire"
  • 3. Evolution of automation Time Degree of automation Electromechanical switch Electronic switch OSS Telephone exchange SDN, NFV, IBN, OMG Brutal Automation Ferocious, Savage, Bloodcurdling Automation
  • 4. Intelligent process automation per McKinsey 1. RPA: software tool that automates routine tasks such as data extraction and cleaning. Robot has user ID and performs rules- based tasks such as accessing programs and systems, performing calculations, creating reports, and checking files. 2. ML: algorithms that identify patterns in structured data, through supervised or unsupervised learning, and provide insights or make predictions. 3. NLP: a way for computers to analyze, understand, and derive meaning from human language.
  • 7. Typical RPA use cases Use case Challenge Outcome Customer Service & Support Desk Average time for 1st line support to execute user admin tasks = 6 mins Task time reduced to 50 seconds with 2 week project saving $1m annually IT & Infrastructure Support Managed service provider Needed a skilled support solution expert for a 24/7 operation. Bot created to handle admin tasks including validation, regular testing, diagnostics and fault remediation. Data Migration & Management 12 employees assigned to manually rekey records from one system to the other. 2 bots created to replace 12 FTE Back Office Administration 300 staff using 18-year old mainframe system and spreadsheets to process queries in 20 minutes Process time reduced to 2 minutes; 12 FTE saving; investment payback 2 months. Connecting Process Islands HR starter/leaver process used disparate systems with email handoff between departments Processing times reduced by 90% Source: Thoughtonomy
  • 8. Examples of RPA in telecoms • Telefonica UK - SIM swaps, credit checks, order processing, customer reassignment, unlatching, porting, ID generation, customer dispute resolution and customer data updates. • AT&T - migration of customer accounts from DirectTV, completing requirement documents for Ethernet services, automation of sales order entry, reconciliation of revenue against assets and inventory. • Deutsche Telekom - 1,000 robots supporting 30m transactions per year and saving 800 FTE • Tele Danmark (TDC) - Customer facing (order handling and customer care), Operations (updating drawings used in network planning), Finance • Elisa - automated pay-TV ordering; payback 3 months.
  • 9. Specialized software versus BPMS vs RPA Source: Blue Prism
  • 10. RPA conclusions • RPA is effective for processes that require predictable and high frequency interactions with multiple applications. • RPA is highly scalable - a robotic workforce can be doubled almost instantly when new products are about to be launched, and then scaled back after the surge. • RPA can be a cost-effective alternative to Business Process Outsourcing of back office functions. • Telecom operations include many mundane and repetitive but essential processes that require multiple systems to be queried and/or updated to complete the task. The tasks must be completed reliably and accurately - a textbook case for RPA.
  • 12. ML vs AI “Machine learning is a subfield of AI, but it’s grown so large and successful that it now eclipses its proud parent.” – Pedro Domingos, The Master Algorithm AI ML Deep learning E.g. pattern recognition, statistical modeling, data mining, predictive analytics, data science, self-organizing systems, etc. E.g. Multilayer Perceptron E.g. Knowledge base
  • 13. Machine Learning types • Supervised learning: in addition to the data inputs we also show the system the desired outputs and ask the system to create its own mapping of inputs to output conditions. • Classification systems e.g. Bayesian networks; Decision trees; Logistic regression; Random Forest; Support Vector Machine. • Regression systems e.g. KNN; Linear regression. A regression algorithm indicates a statistical relationship between two, or more, variables (e.g. temperature and noise). • Unsupervised learning: system tries to find patterns in the input data without knowing of any specific output conditions of the data. Clustering techniques are normally applied to unsupervised learning systems. Examples of clustering algorithms include Apriori; Distribution-based (Gaussian mixture models); K-Means.
  • 14. Why the resurgence of interest in ML? • Breakthroughs in neural network theory around 2006 • Improvements in computing capacity: x86 CPUs, GPUs, FPGAs and custom ASICs designed specifically for ML e.g. Google's Tensor Processing Unit (TPU) and associated Tensor Flow software libraries. • The cloud makes computing capacity highly available and cheap. • Massive data sets: online photos, email, video, gaming, search, messaging, mapping and shopping are fertile ground for ML. • Success stories: AlphaGo, Google Pixel Buds, image recognition, lip reading, etc.
  • 15. Adoption of ML in telecoms 29% 31% 22% 18% Are you deploying AI? We're taking a wait and see approach We have started proofs of concept We are working with suppliers to build AI intro some products and services We have built some internal AI expertise and are incoporating it into product and service roadmaps Source: TMForum, 2017
  • 16. Potential use cases for ML • Customer facing • chatbots • contact center optimization and compliance • sentiment analysis • churn prediction and prevention • fraud detection and prevention • Networking • anomaly detection for OAM&P (operations, administration, maintenance and provisioning) • security – e.g., threat detection • automated resolution of trouble tickets • prediction of network faults • performance monitoring and optimization • route optimization – e.g. in IP transport or SDN switching • traffic identification – e.g. for policy enforcement • SLA monitoring and enforcement
  • 17. AI use cases in telecoms Source: TMForum, 2017
  • 18. ML and analytics for network management Source: TMForum, 2017
  • 19. ML in network management • Vodafone trialed machine learning in a Centralized Self- Organizing Network to identify the optimal settings to deliver VoLTE and to predict traffic hotspots. • KDDI developed an AI-based monitoring system to predict anomalies caused by hardware and software. • Zhejiang Telecom implemented AI engine for route optimization, capacity planning, traffic prediction and dynamic optimization of the network.
  • 20. Barriers to ML in networking • Network engineers typically don’t have backgrounds that include the kinds of mathematical training and experience that are essential in ML. • Lack of good data sets: most data sets are noisy, unnormalized. Most data sets are proprietary. There is no standard way to integrate network data with other data sources like CPU usage, memory, etc. • No useful “Theory of Networking”
  • 21. ML in DT’s RT-NSM Architecture Billions of events Entropy filtering, deep learning neural networks Filter Event categorization Advanced clustering: K-Means, k-NN, mixture models, CURE, deep learning, etc. Recomm endation engine Operations & optimization engine based on deep learning algorithms Reinforcement learning Continuous improvement of operational performance
  • 22. ML code is only a small part of the job Source: Hidden Technical Debt in Machine Learning Systems, Sculley et al.
  • 25. Telecom Infra Project AI/ML working group • Will focus on application of machine-based decisioning and auto- remediation to help carriers keep pace with the growth in network size, traffic volume and service complexity. • Three work streams: • ML-based network operations, optimization and planning to enhance intelligence in network operations areas through, for example, predictive maintenance and dynamic resource allocation • Customer behavior-driven service optimization to enhance the overall customer experience, particularly for bandwidth-intensive, latency-sensitive and/or data-heavy applications • Multi-Vendor ML-AI Data Exchange Formats to ensure the development of generalized ML models that are applicable across the industry.
  • 27. The Route To Automation Traditional networking Cloud networking Manual Software-driven Delivery time 10-20 days (enterprise service) Delivery time 5-10 minutes Static (>1 year contracts) Dynamic (on demand services) Source: Colt
  • 28. Automation end game If we virtualize all our networks functions and use big data analytics and deep learning algorithms to run the operations … … the entire business can be managed by just one monkey … … plus a second monkey to look at the PowerPoint slides of the first monkey.
  • 29. Paradox of Automation • “The more advanced a control system is, so the more crucial may be the contribution of the human operator.” - Dr. Lisanne Bainbridge, Department of Psychology, University College London. • “We appear to be locked into a cycle in which automation begets the erosion of skills or the lack of skills in the first place and this then begets more automation” - William Langewiesche, ex professional airplane pilot. • “To err is human, to really foul things up requires a computer.” - William E. Vaughan, newspaper columnist.
  • 30. Q&A