The AI Magic in Telecom — an applied AI use case
Written by Deepak Ranjan
“Everything invented in the past 150 years will be reinvented using AI within the next 15 years.” — Randy Dean
With increasing computing power, cheaper storage cost — data has become the new oil. With technology augmentation with smart devices, faster 5G network and expansive interconnectivity among devices with IOT — mobile devices are a percipient decipherer of user’s activities and behavior. This makes the telecom industry “an oil well which is never going to dry”. A telecom network allows an unprecedented volume of data to flow through their systems and has extraordinary use cases.
The worldwide telecoms sector is anticipated to invest $36.7 billion on AI software per year by 2025. The graph below illustrates that network operations monitoring and management will be a significant area of AI adoption.
Although AI has plentiful business ideas specific to the telecom industry and they are cogent for any Communication Service Providers(CSPs) around the world, Circles.Life is not an exception. Being on a mission to give power back to customers — that’s where the true motivation sprawls and steer the organization to go above and beyond. We will have an incisive picture about a few of them in this article.
Network Optimization
Self Optimizing Networks(SONs) are the next big thing Telco's are talking about, allowing telco operators the ability to boost network performance and quality. Forecasting network anomalies based on the data pattern empowers firms to fix the snag before they have a detrimental impact on customers. Machine learning aids Telco's in optimizing network performance KPIs and KQIs in order to provide real-time network views. Seamless network orchestration across the network is another AI enabled feature firms are interested in. Advanced AI algorithms can act as safeguard against sudden network traffic overload and provide means to tackle this problem without any human intervention. AI backed Network Function Virtualization(NFV) abolishes the need of custom hardware with virtual machines, it can automatically channelize the traffic without any external interference. At Circles.Life, having a loosely coupled ecosystem components in place, seamless customer experience of our platform comes with equal significance along with network optimization for telco usage. Building software systems which can auto forecast load, deploy remedial steps, intelligent automation to minimize manual involvement is our utmost precedence.
FACT: As per a FiorMarket study, 63.5% telecommunication organizations are making new investments for AI systems and 31.5% are primarily focusing on leveraging existing infrastructure.
Virtual Assistant
Telecom is a 24*7 essential services. This necessitates round the clock customer service which comes with a massive financial and resource overload. Current scenario where the majority of Telco's are struggling to manage their finances, many more firms are opting for AI enabled VAs in order to save some bucks and better human resource utilization. A good portion of customer requests can be classified based on their similarity and this is where AI backed VAs come handy and work as per trained SOPs — it helps in simplifying front line support, solve complex issues quicker.
FACT: As per Gartner 25% of digitals workers will use virtual assistants on a daily basis by 2021.
At Circle.Life, currently we are serving an ever growing multilingual customer base in 5 countries and growing exponentially YoY. Natural Language Processing(NLP) based VAs supporting the world’s major languages will be the need of the future. Telco service offering as a platform where our app features would be growing multifold, customer service will remain the first touch point for not so tech-savvy customers. It’s really engrossing to visualize a VA mapping the customer’s text/voice message to the system APIs and serve them without any external interference. Indeed, a moment of sheer customer delight!
Business Decisions
Although there can be number of business scenarios possible in a digital space to derive insights from user’s behavioral data, keeping the scope of this article in mind, we would be talking about couple of exciting use case here:
Churn Prediction: Due to easy mobile number portability and aggressive price war in the telco industry, customer churn prediction has become one of the fundamental business problems even for the most successful CSPs in the world. It is unopposed accepted that doing business with existing customers is more economical and profitable than acquiring new ones — this elucidates the importance of predicting churning. With early detection of churning customers, customized strategies can be modelled for individuals and it also helps in strategizing the business by digging down the root cause.
Telco Credit Score: We are moving towards a future where telco tech giants would be offering AI backed alternate credit scores as one of the most sought non telco offerings to the world. If we factor in the amount of user’ behavioral and historical data a telco firm collects during a lifetime of a customer. The insights from such data would be nearly unparalleled & exceptional. A sound multi class classification model on such a dataset is something to count on!
Robotics Process Automation
CSPs engage in millions of transactions every day with their customers. As a result, it’s vulnerable to human error and communication stumbling blocks. That’s when the efficiency of RPA may be put to work in the telecom business. It helps in managing immense rule-based but repetitive tasks which is an integral part of any enterprise. At Circles.Life, where a true digital experience is our top priority, business flows like workforce management, order fulfillment, data entry and billing flows are few of the areas touched upon by RPA robots. In the telco industry, RPA may improve customer service, data quality, cost savings, and operational efficiency. It aids in the dissemination and management of periodic reports.
FACT: MRFR predicts that by 2023 global AU telco market will reach $1 billion with 32% CAGR during 2018–2023.
Conclusion: The importance of Artificial intelligence in univocally accepted and telco industry is not an exception. It has been welcomed with open arms by the CSPs and has already been used in major business flows. It still has many open prospects all around the system, few of them are in a very naïve phase and need extensive research.
Deepak Ranjan is working as a Telco BSS SME in Circles.Life & comes with 7+ years of experience in this domain. He is having 30+ academic projects and case studies in AI/ML and holds a PGD degree from IIIT Bangalore. He is currently pursuing his Masters(Research) from LJMU, UK. He is inspired to work as a ML engineer in future, role having fusion of Software Engineering and Machine Learning excites him.
