Shankar Arumugavelu is what you might call a Verizon lifer. He was a director at telecom GTE when Bell Atlantic acquired it in 2000 to form Verizon. Today he\u2019s SVP and global CIO of Verizon, where he\u2019s helping to drive the company\u2019s adoption of emerging technologies like AI and machine learning in service of creating competitive advantage and improving customer experience.\n\u201cAs we look at emerging technologies, AI is a big area of focus,\u201d Arumugavelu says. \u201cYou have disciplines within AI as well, whether it\u2019s NLP or computer vision, robotic process automation, cognitive decisioning, etc. We have work going on across every single one of those disciplines to see how we can leverage that to drive a competitive advantage.\u201d\nArumugavelu and his team evaluate technologies based on multiple criteria, but the ability to drive operational efficiency and to deliver a differentiated customer experience are two of the most important factors.\n\u201cWhen we talk AI and machine learning, these are technologies that have been there for many, many years. It\u2019s just that now the time has come,\u201d he says.\nData is the raw material that powers all these technologies, and Arumugavelu says Verizon has \u201cno paucity\u201d of it. Along with the growing volumes of data, there\u2019s been a steady decrease in the cost of compute, greater accessibility of AI and machine learning research and algorithms, and increasing availability of tools to help democratize data.\n\u201cThe four factors put together are giving us an opportune moment to really capitalize on these emerging technologies,\u201d Arumugavelu says.\nNLP streamlines customer connections\nNatural language processing (NLP) is a key example of an emerging technology whose time has come.\n In 2016, Verizon decided to add an NLP-based chatbot to its mobile app. The company had already successfully experimented with an internal chatbot service for its techs called IVAPP Buddy, which gave it the experience it needed to tackle a customer-facing app.\nThe first version of the chatbot for the My Verizon app was relatively rudimentary. Customers could ask a question and the chatbot would come back with an answer, based on a list of frequently asked questions. The customer response showed promise and the team decided it couldn\u2019t stop there.\n\u201cThe technology had matured to the point where our goal now has to be to have this virtual assistant support multi-minute conversation completely in an automated manner,\u201d Arumugavelu says. \u201cThat\u2019s only possible if the bot is going to be able to maintain the context of the conversation. For instance, if the customer is asking question after question, and a new question relates to something that was asked several steps before, the bot should still be able to understand the context of that conversation.\u201d\nFrom there, the team has continued to iterate on the bot\u2019s capabilities. They were satisfied the virtual assistant could successfully help customers and resolve most issues, but they also saw potential in a secondary role for the chatbot: to assist a live chat agent if a customer decides the virtual assistant can\u2019t address their issue. The next iteration was building a conversational user interface based on voice, an interactive voice response (IVR) system\n\u201cUltimately, you have the same corpus that\u2019s driving this multimodal experience for customers, irrespective of whether they\u2019re coming through a chat or through the IVR,\u201d Arumugavelu says. \u201cIt gave us an opportunity to reflect our brand and persona across all the customer touch points and deliver that consistent experience to our customers.\u201d\nArumugavelu explains that customer satisfaction has been good, and the technology has also been successful in containing incoming calls and chats within the automated platform.\n\u201cI\u2019m using the term \u2018containment\u2019 deliberately, because for a long time, the efficacy of these tools was just measured in terms of call deflections,\u201d he says. \u201cBut deflecting is not the point here. If a customer comes to interact with the channel, how are we giving the customers the kind of experience that enables them to complete whatever they wanted to do in that channel without having to go anywhere else?\u201d\nThe lesson here is to measure the right things. Arumugavelu says that\u2019s the key to understanding the real impact of emerging technologies on your business.\n\u201cLook at it from the customer perspective: Were they able to get the things done that they came here to do in an automated manner? It\u2019s not that they hung up from that session, but we also did not see a call back or the customer going to some other assistant channel to ask the same question and get clarification.\u201d\nDigital twins boost network reliability\nEmerging technologies also have a role to play in helping Verizon plan, build, and maintain its network of wireless towers and global wireline network. The company is using digital twins, which serve as a bridge between the physical and digital domains, providing a real-time virtual representation of physical objects and processes, to gain new visibility and insight into those networks.\n\u201cWhen you plan a network, you engineer the network, you go construct it in the physical world, and then you have to figure out if the engineered view and the as-built view are really one and the same,\u201d Arumugavelu says. \u201cHow many antennas do I have? What is the tilt. There are hundreds and thousands of different parameters that go with each of these cell towers.\u201d\nVerizon is leveraging drone imagery and computer vision to understand the configuration of its cell sites and then comparing the results with the engineered view to determine whether the two are in sync. If there\u2019s variation, Verizon can make changes at the cell site to bring it back in line. That\u2019s critical, he says, because it can be difficult for a human to analyze all the data from a particular site if it starts experiencing problems.\n\u201cThis is another area where machine learning is applied to be able to predict, analyze the network performance data, the alerts, and also predict what the alert\u2019s impact would be and figure out the root cause for the incident in the first place,\u201d Arumugavelu says. \u201cIf we don\u2019t get that right at the beginning, the second order or third order effects of that are big.\u201d\nThe digital twins also help Verizon optimize its network performance and preventative maintenance schedule. Any time a change is made to its network, it can see the downstream effects.\n\u201cWe are taking action before something goes wrong, and that ultimately translates to the reliability of our network,\u201d Arumugavelu says. \u201cThat\u2019s a really big priority for us. It\u2019s our crown jewel. Customers come to us for our network reliability, and this plays a big part in ensuring that we are able to deliver on the promise.\u201d\nSolving the business problem \nAs a CIO, the million-dollar question comes down to which transformative project you should focus on. Arumugavelu says that he always starts with understanding the business and its needs.\n\u201cAre we solving the right kind of problems for the business versus a bunch of science experiments? It starts with that,\u201d he says.\nEven then, he says prioritizing the use cases is key. Sit with business stakeholders, understand their needs, and help them see the art of the possible with technology.\nBy way of example, Arumugavelu notes that at the beginning of the COVID-19 pandemic, as many people started working from home and schools went virtual, the reliability of its broadband connectivity became essential to many of its customers. It added many technicians to aid with installs, repairs, and so forth. But the business was wrestling with an important issue: How could it maintain employee and customer safety in those circumstances? Could it do most of the required work without a technician ever stepping foot in a customer home?\n\u201cThat led to the idea of using augmented reality to provide remote visual assistance to our customers without necessarily having the technician go into the customer\u2019s home and put himself or herself, and also the customer, in jeopardy as well,\u201d Arumugavelu says.\nThe technician could connect with the customer via mobile app and the customer could use the rear-view camera on their smart phone to allow the technician to see the customer\u2019s equipment. Verizon has since brought that capability into its call centers, so agents can now help many customers solve their problems without sending out a truck.\n\u201cIt\u2019s a win-win both from a company and a customer perspective,\u201d Arumugavelu says.