Big Questions Answered helps us understand important CVS Health initiatives by taking a closer look at new products, powerful innovations and the big changes the company is making to achieve its strategic imperatives and build a world of health around every consumer. The company's senior leaders answer big questions from host Matt McGuire.
Matt McGuire
Artificial Intelligence has become an incredible tool in modern health care. It helps industry professionals as they analyze lab results, predict patient risk and make clinical decisions.
At CVS Health, we’re integrating AI across the company, from our clinics to our digital chatbots. On this episode of Big Questions Answered, we explore how AI is harnessing data to help patients improve their health outcomes, provide call center colleagues with real-time insights and drive measurable improvements in customer loyalty and satisfaction.
Welcome to Big Questions Answered, a podcast that helps us understand the important initiatives at CVS Health. I’m Matt McGuire from the Enterprise Communications team. I’ll be your host as we dive into the latest developments, breakthroughs and innovations that are helping us achieve our ambition to become America’s most trusted health care company. Thanks for joining me today as we get our big questions answered.
Josh Weiner is our guest this month. He’s the Senior Vice President of Consumer Engagement & Analytics with the Digital, Data, Analytics and Technology organization within CVS Health.
Josh and his team help engage, grow and retain our consumers by building and implementing personalization algorithms across channels. They also help our colleagues deliver better customer service by building decision support tools.
Josh, thanks for stopping by.
Josh Weiner
Thanks, Matt. Grateful to be here.
Matt McGuire
So, Josh, my understanding of the work that you and your team are doing is that it’s designed to help people make smart decisions. And, to me, that’s fascinating. From what I’ve read, you’re using AI and analytics to do this. Can you tell me how you’re looking at large sets of data to understand people’s behavior, their preferences and their health trends? And then using what you’ve learned to improve their health outcomes or make them aware of relevant retail deals?
Josh Weiner
Matt, that's a great question. We have a three-step approach to determine the most effective way to be helpful. The three steps are: One, at the consumer level, how do we help, what's the challenge you're facing or what's your next best action as applicable? Two is what do we actually do to be helpful, which includes understanding your barrier or issue as well as among the interventions or tools in our toolbox, what makes the most sense for you. And three is when, where and how do we communicate this, taking advantage of our vast ecosystem of digital and physical engagement channels? So, let me double-click on each of those three steps with an example.
So, step one is how can we be helpful? What behavior should we change? In the context of health care, this typically fits into a couple of categories. It's benefit navigation. What do my benefits mean? Help me pick a provider. Let me figure out how to save the most money on this procedure. It also fits into wellness. What sort of screenings should I get? What sort of vaccinations am I eligible for and should I get? And then lastly, in many segments, there's a servicing component. I need to pay a bill, I need to update my product, I need to make a change. So, step one at the mat level, what's going on and an always on point of view of how to be most useful.
Step two, what action's most likely to work? Given our ecosystem of levers and tools, this can be reminders, this can be educational material, this can be connecting you to a nurse or a doctor. This can be summarizing changes to your benefit design. This can also, of course, be taking an action on your behalf. It's often a function of your preferences, your specific needs and challenges mapped to where you are in your health journey or the product you purchase from us.
And then, lastly, when and where and how do we actually engage? We have our stores, we have our contact centers, we also have outbound channels like e-mail messaging and of course our awesome app and digital experiences. So, it's through this three-step approach that we make the most of our historical understanding of who you are and the challenges you're facing, use that to assign an intervention or action that makes the most sense for you and then actually go ahead and deliver on that in the way and channel that meets you where you are.
Matt McGuire
So, that’s impressive and in-depth work that you and your team are doing. And with all of that in mind, how are you using AI to take personalization to the next level?
Josh Weiner
It's a great question, and I appreciate that the framework in of itself isn't all that novel and it's something we've been using for quite some time. AI has transformed our ability to do this well. And let me explain that going through each step.
So, we think about the first step, which is how can we be helpful? Pre AI, we were limited to, one, working off of data sets that were relatively well structured. So that often means kind of claims data. And due to the sheer feasibility constraints of the analysis, often grouping consumers in, say, 20 to 50 segments just to be pragmatic about the subsequent workflow. AI, one, allows us to now summarize unstructured data incredibly well. So, think about medical records, historical interaction data, contact center history, provider contacts, etc. And, two, allows us to now finally do this at the individual level and move from cluster level personalization to individualization. So that's part one.
And then in the context of how we actually engage, we can take e-mail as an example. Pre-AI, once again, for practicality reasons, we're often forced to at most have, say, four variants of an e-mail — and maybe within that a couple different subject lines.
Finally, with AI to the extent it makes sense, every single person can have their own individual communication, and that's solely because AI enables at-scale personalization — subject lines, content, creative and a subsequent workflow that allows us to execute on that, revolutionizing marketing automation.
Matt McGuire
Those are some notable transformations, and that definitely outlines how AI is helping our colleagues help customers, patients and members. But as we think about a percentage of the people we serve, those who are a bit older, when we combine them with the sensitivity of health care, it starts to become clear why a decent sized group of the people we serve prefer to make calls on their phone. And that moves away from much of what you were just talking about. What are we doing to transform the experience of a phone call?
Josh Weiner
Another great question. We apply AI throughout the lifecycle of a call to improve contact center experiences. So, there are four applications. Starting with pre-call, it's through AI that we are more accurately able to predict why you're likely to call, and we do two things with it. One, when it makes sense, we now route you to an agent that's better suited to handle your issue. And two, when you're connected to that agent, on their desktop, their content is pre-populated based on our understanding of your call reason, which allows them to have a more productive and effective conversation. So that's smart call routing and use case one pre call. Use case two is agent assist.
Across our contact centers, our agents now have AI built into their application to help them answer complex questions more accurately and more quickly.
So, historically, if we take Aetna, for example, a question pertaining to say what's the cost of a colonoscopy would be incredibly difficult to answer well and certainly quickly. It would often require navigating between 4-5 different applications, mapping and understanding of your benefits, to what does that service mean from a CPT code perspective, to what are the prices associated with the physician you have. Through AI, we can do this automatically and we present the capability to our contact center representatives on their homepage. So, what previously, say, took 5 minutes of research, now gets done in less than 20 seconds. So, that's agent assist.
As we work through the lifecycle of the call, we now have call summarization. Historically, our agents would spend time both during the call and after the call documenting what happened. This is no longer necessary. Through AI, we're able to summarize the calls. They're more consistent, they're more comprehensive. Our representatives take a look at the summary, make sure it's correct, add any additional information and press submit. It's much faster, it's much more effective and it allows us to now share these summaries across the ecosystem, which in turn improves consumer experiences. We call this call memories as we roll this out throughout the company.
And then lastly, we think about post-call — performance management, learning and development training. Historically, the best any company could do would be to listen in on calls, to randomly sample calls and to use surveys to figure out what's going well and where there are areas for improvement to provide coaching for contact center colleagues. AI enables us to do this for every single call, for every single employee, and increasingly opens the door to real-time coaching and feedback based on the actual interaction.
Matt McGuire
That’s great. I mean, if this improves the consumer experience and the colleague experience, both from the perspective of call center employees and their managers, it’s gotta be viewed as a win-win, right?
Josh Weiner
One hundred percent.
Matt McGuire
Yeah.
Josh Weiner
Through AI, we've been able to take some of the more cumbersome and grueling aspects of work out of one's responsibilities and in turn are allowing people to spend more time really focused on doing what they do best and prioritizing our consumers. So, that applies in the case of our contact center colleagues, it applies in the case of our clinical colleagues and many of our colleagues cross various functions. So, it's definitely a win-win. I'd also say that we're fortunate to work in a space where what's generally best for our consumers is also best for our company. In the context of, you know, behavior change programs and personalization only help people get the care they need or make better use of their benefits. It not only benefits them and helps them save money out of pocket, but by improving their health trajectory, it's good for our company as well.
Matt McGuire
And to build on what you just mentioned — what’s good for the company — I’m curious, in what additional ways does AI help CVS Health?
Josh Weiner
Matt, when we help consumers, we improve company performance. There's a well established relationship between consumer satisfaction, Net Promoter Score and growth — growth through retention and growth through increased sales. When we improve the colleague experience, we also help our consumers and help the company, so it's really as simple as staying laser focused on what's best for our consumers and through that we build trust, we build engagement and everything else follows.
In the context of health care, we're often doing this by also helping people get care, get the care they need more quickly, more effectively, more proactively. A large portion of our engagement is tied to prevention and wellness. And as a health plan, when we help people improve their health trajectory, that also flows through into health plan economics in a favorable way.
Matt McGuire
So, Josh, we’ve talked a lot about how things are and how they’ve been going. Let’s take a look at the rest of 2026 and a bit into 2027. What’s on the horizon for you and your team?
Josh Weiner
Perhaps I'll just share something that's really exciting. And that is our company's advancements in messaging. Over the last couple years and moving forward, standard text messaging is being systematically upgraded to a more advanced protocol that looks and feels a lot more like WhatsApp, if you or anyone listening uses WhatsApp. And when you combine that with the personalization capabilities and kind of AI chat related capabilities that we're making so much progress against, we're going to really unlock an entirely new channel for consumers, and a wonderful channel that can coexist with our rich digital ecosystem and present a more convenient alternative to calling and talking with a human.
Increasingly, I expect consumers to appreciate what they can do simply through texting CVS or Aetna natively and iMessage. And CVS is already the farthest along across industry with respect to our RCS or rich communication messaging capabilities and we're going to continue investing here to meet consumers where they are and emphasize convenience and simplicity.
Matt McGuire
I am looking forward to that. Josh, this has been an extremely interesting conversation. I've had a great time talking with you. Thank you very much for stopping by.
Josh Weiner
Thanks, Matt. Have a good one.
Matt McGuire
And a big thanks to you for tuning in to this episode. Until next time, I'm Matt McGuire. I look forward to joining you again to get more big questions answered.