While consumer habits and technologies have changed rapidly, the technology function within organizations has tended to be slower to adapt. That realization—and the desire to create experiences that its customers expect—inspired Verizon Consumer and Business Group (Verizon Consumer Group) to make sweeping changes to its technology organization, from a focus on supporting channels and platforms to supporting complete customer journeys. This shift required a broad set of architectural and operational changes, both within and outside of technology, to take advantage of new technologies, such as gen AI, and develop better products and services more quickly. McKinsey’s Mark McMillan sat down with Vivek Gurumurthy, senior vice president and CIO of Verizon Consumer and Business Group, to discuss the challenges, decisions, and learnings from his group’s ongoing transformation. What follows are edited highlights from their conversation.
The case for change: Catching up to the customer
Mark McMillan: How has Verizon Consumer Group’s technology strategy shifted over the past few years?
Vivek Gurumurthy: Customer needs have evolved significantly over the years, from wanting to conduct a transaction on a mobile device in a store to being on the go with a digital experience at their fingertips. When you think about what the customer is looking for, it’s a journey that they can start and stop anywhere—a seamless digital and physical interaction. It’s a totally personalized, context-centric experience.
But the technology landscape supporting the business hasn’t really evolved or changed since the introduction of the wireless business about two decades ago. So we really had to step back and take a hard look at the technology footprint and ask ourselves, “How do we invest and reinvent the technology architecture to support the changing needs of the customer?”
We looked at our systems and saw that every channel was supported by siloed architecture. It wasn’t designed for customer experiences; it was designed for customer transactions, which were very channel-centric.
So we embarked on this journey about four years ago to change our systems landscape from one focused on channels to one focused on the entire customer journey. That was our North Star.
Mark McMillan: Can you say more about why customer-centricity is so important to the Verizon Consumer Group?
Vivek Gurumurthy: The telecom space is a saturated market right now. You have three major providers, and then you have cable companies coming in as disruptors. Everybody in the country now has a cell phone, so it’s a game of moving customers across these providers. Stickiness has repeatedly favored companies that provide the right experience at the right cost. As a premium provider, it is important that we focus on, and invest in, delivering value, while also delivering the right experience.
The right experience is not just a simple digital transaction that takes less than five seconds. It’s more about the white-glove treatment, knowing the customer, the context of the transaction and the conversation, and presenting the right proposition at the right time. So essentially, we really had to think about how to transform from a traditional telco into an AI-led one.
Building technology foundations to innovate the customer experience and create value
Mark McMillan: Given these shifts that you’re describing and technology’s increasingly important role for the business, what are some of the innovations you delivered to support the business?
Vivek Gurumurthy: As you know, COVID-19 was a global accelerator for digital transformation, and it was no different for us. Overnight, we had to transform the business from a physical environment to a completely remote one. Supporting the frontline employees and call centers was easy, because we were already globally enabled. But supporting customers in a remote environment, where they could do everything digitally and get their devices and connections without talking to an employee, required serious innovation. We installed lockers in our stores and created in-store pickups, which kept our volume up, kept our customers going, and kept the device innovations coming.
It was important that we not only enabled digital but also enabled these innovations using data and the AI we had at the time. We started looking at the right offers to provide customers at various touchpoints to either retain or upsell them. This focus had been almost ten years in the works for us. Post-COVID-19, we were able to position ourselves very well for the next cycle of AI innovation. We invested in an AI contact center, which enabled our customers to transact through chat bots or conversational AI. We also enabled our reps to gain instant access to information about the customer, the interaction, and the context. Most importantly, we had the best-in-class voice recognition and natural-language recognition engine at the time.
These critical elements were foundational for us and put us in a beautiful place to capitalize on generative AI. We got in early during the innovation cycle. Our AI teams built models and created platforms that helped our reps support our customers with the right information, and we augmented our architecture to support generative AI. We now have more than 40,000 reps in our call centers using active generative AI–aided tools to support our customers on all kinds of transactions.
The big technology shifts needed to enable innovation
Mark McMillan: What were the big shifts you made to take the organization through to deliver on the innovation you described?
Vivek Gurumurthy: For us to capitalize on AI and the shifts in technology, we had to fundamentally reimagine our architecture. We were primarily on-premises, which doesn’t allow for rapid adoption of innovation and scale for digital loads. So the first thing we did was move all our critical loads to the public cloud.
Step two involved data and algorithms. When we started on AI ten years ago, we were just getting the data together, organizing it in a structured format, and feeding it to the algorithms. But now, the algorithms are deciding what data they need to provide insights and getting it themselves. So that’s a huge shift in how we organize data, algorithms, and models as they’re being used in real time during customer conversations and transactions, and not just in analytics for decision making.
The third step was ensuring that our people were ready for all this, investing in them and making sure they were upskilled to complete their knowledge transformation from legacy systems to the public cloud to AI-enabled technologies.
The last and most-critical step was ensuring we had the right ways of working to deliver this technology to the business and capture the value. That meant moving away from a traditional waterfall engineering model to a completely product-based, agile delivery model, where we are cocreating with teams from the business side of the house. This approach helps ensure our solutions are delivered and adopted to realize the value. This cannot be overstated. If you build fancy tools that aren’t being used for the right purpose, there is no value capture.
Crucial to this last step was a program called Marvel, which completely transformed our engineering practices to support the agile transformation and product model. Our engineering practices had to be retooled from completely on-premises to a cloud-based, continuous integration and continuous delivery (CI/CD) methodology.
Mark McMillan: What was the Marvel program, and how did it help with the transformation of your engineering practices?
Vivek Gurumurthy: “Marvel” was a name coined by the team for the modernization of our agile and reliability practices. We modernized our architecture and our ways of working with agile practices and changed our reliability practices to a new site-reliability engineering (SRE) model.
Building this infrastructure to support an SRE model required a complete reimagining and rethinking of how our developers used tools to build systems and how our testers used automation to quickly test thousands of scenarios to validate the security, compliance, and quality of the code. Previously, we delivered about ten releases a year, but now we deliver every two weeks throughout the year, even in the middle of the day.
Marvel also provided us with tools to test functionality, and automation systems to test data. We also built a CI/CD pipeline with real-time code validation, verification, and compliance tools.
One hidden value of Marvel was answering a question we kept asking ourselves: “How can we improve the quality of our digital experience?” As much as Marvel allowed us to move faster and build new tools, it also improved the quality of delivery, since we could now do test deployments, limited releases, and A/B testing in real time. We could also incorporate feedback from users to improve quality before rolling out at scale. Marvel’s biggest benefit was both quality and speed to market.
More than just tech: Rewiring how we work inside technology and with the business
Mark McMillan: What are the biggest changes you made to your operating model and how your business and technology teams partner with each other?
Vivek Gurumurthy: The toughest part of any transformation is the people transformation. A year before this transformation, we were strictly organized by applications, so employees would identify themselves, their roles, and their careers with the applications and platforms they supported. In order to stitch together an experience, the architecture team would have to build solutions and get teams engaged from all these applications, which would then deliver the capability.
We shifted the entire paradigm. Even the technology developers and the development teams in the technology organization have to be oriented toward customer experience, not the applications. That meant a complete rewiring of the organization, bringing together people who had never been on the same team. It also meant teaching managers new skills, like product-management fundamentals, since they were only adept at managing their particular platform in a delivery cycle or phase of a delivery cycle.
Now, the product team is a full-stack one, with every manager responsible for understanding the full delivery cycle, from design to implementation. We’ve changed the way SRE works and, most importantly, how the interactions with the business happen, who faces the business, and who runs the platforms on the back end.
Driving change management without losing momentum was probably the most important part of this transformation, because you had to convince employees that they need to think differently about their careers, and they have to learn new skills across multiple applications to deliver the product and experience, not just the platform or application they’re developing.
This was a two-year journey we accomplished in phases. But when we did the cutover, nearly 3,000 employees across the organization had new managers overnight. The transformation continues to evolve, and I would say we are now in a reasonably mature state.
Mark McMillan: How did you navigate that change with the leaders on the business side and bring them along on the journey?
Vivek Gurumurthy: The technology transformation was a catalyst for major organizational changes in the company. One was that our business-facing, frontline-facing, and digital teams had to rethink how they were organized within their own structures.
At the time, there was one team providing stories and requirements for the digital channel, another one doing the same for the retail stores, and another for the call centers. But then they were all forced to think about journeys that touch all channels during an experience. That was a catalyst for changing the operational support structure and putting them together into customer-experience product teams, which now serve as single entities to drive experiences across all channels.
The second part of the transformation was a huge drive to change the way our HR team thought about people and their careers. This forced our HR team to think about technology, about experience-based organizing, and about titles that more accurately reflected new employee roles.
The final part of the transformation involved changing how the finance team thought about technology capital and operational expenditures, as well as how they funded all the different product teams.
The last thing we did was change how we did prioritizations. What used to be a once- or twice-a-year prioritization of programs is now a regular monthly or quarterly prioritization. The planning sessions allow us to prioritize the right work for the right products and give product teams autonomy to decide what their backlog looks like.
Taking a step back: Transformation lessons learned
Mark McMillan: What are the top lessons learned you can share with other tech executives on their transformation journey?
Vivek Gurumurthy: The most important thing is to always focus on value for the business. If you are able to deliver the value the business needs, that gives you the space to innovate and attracts the necessary investment to change other aspects of the business.
Beyond that, I would recommend, number one, that you make sure you have the sponsorship of the leadership. Sponsorship of the technology leadership is crucial, but sponsorship of the business leadership is also essential, because there will be some hard times when you need to make changes to accommodate the plan.
Number two, get buy-in from your employees. The employees in our organization were craving this change, and when I had that support, it was easier for me to walk them through the transformation.
The third thing is ensuring that the business is ready for a new operating model. The tech side is the catalyst for the change, but the business side has to be ready to receive the change.
Lastly, do not start the transformation without knowing the end-state architecture. Traditionally, architecture evolved to mirror legacy organizations. But in this case, you have to put the architecture in place and then build the organization to support that architecture. Take the time to plan the end state—not just how the systems will look, but how the operating model will look and how the ways of working will change. Highlight and document all of those, get the stakeholders engaged, and help them through the process. Once you do this, it becomes an easier process.
