David Arturi: Today I have the pleasure of speaking with Dr. Tiffany Perkins Munn, managing director and Head of Data and Analytics for JPMorgan Chase, Marketing. Do you have any best practices or advice when it comes to privacy and data security?
Tiffany Perkins-Munn: In financial services, and I know that this isn’t true for all other agencies, and other organizations, but they may have other requirements like this. There’s this idea that you have to comply with regulations right. You have to adhere strictly to financial regulations. We have things like GDPR, CCPA, Industry specific laws. So, implementing a robust data governance framework to ensure that you are compliant across all AI initiatives is the very 1st step that anybody should take when they’re thinking about data privacy, and I am suggesting that you understand what that means for your organization, or for your industry or for your business. I just gave examples of what that means for us. But that kind of robust data governance framework is really important. Also, as we’ve kind of alluded to this point as well, there’s data minimization. Right. You’re only going to collect and use the data that’s necessary for your specific purposes. And then you are going to regularly audit and purge unnecessary data, because sometimes people bring in data, it gets old and stale, and it’s like a year old, and you know, people, in the age of apple and technology and immediate access to information, things are updating quickly. Right? So last year’s information, last month’s information, yesterday’s information is often old. I think that you have to, for the sake of the customer, you have to think about encryption and security. You want to use strong encryption for data, either data that’s at rest or data that’s in transit being ported from across Via API, for example. And then have vigorous access controls like who is allowed to access this data, and how? What are they allowed to do with it? And how do you authenticate right to make sure that it is the authentic source of the information that you need. I think consent management is important, clear, informed consent for data collection and use that will ensure its privacy and easy opt out mechanisms for consumers. You’ve been on a site where it’s like I’d like to unsubscribe, and then at the bottom, it’s like click here to unsubscribe, and then you go there, and you try to unsubscribe. You can’t even get. You can’t get off. You can’t get away. And then through some loopholes they can keep sending you stuff. That makes customers not really trust you. So, you want to make sure that you’re providing an easy opt out mechanism. And oh, this is important. Data anonymization right. You want to use techniques like tokenization or differential privacy. You want to make sure that you’re protecting individual entities while still deriving the insights. And I think that’s important in that partnership with consumers to say we are not only is your data confidential, but it’s also anonymous. We are not attaching any of the information that we have to you and sharing it out publicly. And that’s key as well.
David Arturi: I think that’s critical, especially when we start bringing AI into the mix, because again, everybody thinks that AI is just taking it and shipping it out and training their models and we’re not going to get into open versus closed source. But certainly, you know, in some cases that is happening. So, I think those are great points. And it’s also interesting, too, because you mentioned, you know, on one hand we’re saying, take these different data sources, but at the same time you’re absolutely right, having too much data isn’t a good thing, right? You don’t want to have duplicates, you want to have just the data you need just what’s important for multiple reasons, right for training the models, but again, also security. That’s a great way to get breached to have data all over the place that’s on like PII, whatever it is. That’s a great way to get preached. It’s kind of this interesting dynamic where it’s take these different sources, but don’t have too much.
Tiffany Perkins-Munn: Be strategic about the sources that you select and then make decisions as you go through the process, you know, prioritize the data you want.
David Arturi: Yeah, absolutely. So, then kind of just moving to the next one. So, once you deem the data is in the right state, what sort of AI initiatives are you seeing to be the most effective in streamlining processes? One specific area I’d love if you could make sure to touch on would be personalization at scale to drive a stronger customer experience.
Tiffany Perkins-Munn: Yeah, personalization at scale is everybody’s favorite topic. Right? It’s like, everybody’s like, how do we personalize? How do we hyper personalize. Because that’s what everybody wants. The interesting conundrum is that people want you to know exactly what they’re thinking, exactly what they want to buy or purchase next, exactly what you know you should be sending them, but they have hesitation in actually sharing their information.
David Arturi: I’m one of those people. Hand-off. I don’t want to tell you anything, but I want you to know.
Tiffany Perkins-Munn: That’s 1 of the reasons I think there’s an opportunity, for like this partnership between the consumer and brands which we don’t have now. You know, and I don’t mean just like Oh, let’s send them something that says they can opt out. I mean a real partnership like this is how we’re using your information, and this is why, when you, you know, go to Amazon to look for something, when you go back, you actually see something similar and then you start to feel weird because you’re like, Oh, my God! I just bought this thing and now they’re stalking me right. But I think a communication strategy with consumers about how that information is used and how it’s helpful for them, because then it gets the right type of recommendation in front of them and I just don’t think as an industry, we’ve built that kind of back and forth with consumers around privacy, security, and how we use their information, for their benefit. But anyway. Back to your question around personalization at scale. I think that once the data is in a suitable state, there are several AI initiatives that can be highly effective, for, like streamlining the processes, enhancing personalization at scale, all towards the goal of improving the customer experience. And I think the most common one that I see is really recommendation systems, right? So, they analyze customer behavior, they analyze purchase history and preferences, with the goal of suggesting relevant products or content, and that really helps to drive engagement and increase conversion. At least when we’re not making the customers feel like that’s creepy, that we know the product or content that they’re interested in. Another one is dynamic pricing, so AI can optimize pricing in real time, based on demand, based on customer segments and other factors and personalizing offers for individual customers, which is important, especially from the pricing perspective. Something that I think I mentioned earlier chat bots, and virtual assistants. Using advanced NLP, natural language processing, to provide personalized customer support to answer queries like figuring out what are the most common questions that that people ask when they call? And how do you guide them through a process like, how do you guide them through the chat bot through a sales process, for example. Another way, predictive analytics. Because that helps you to anticipate the customer’s needs right. So, that’s a key piece of personalization. What does the customer need? Are they a potential churn risk? What are some of the proactive measures that we could implement to improve customer satisfaction? Which also makes me think of segmentation. Because different segments of customers tend to have different levels of satisfaction depending upon knowing what their needs are within a segment. But AI can actually create more granular and dynamic customer segments. I think that firms customize segmentation in a way that is actually not customized. It’s like very sort of narrowly focused and binary in certain ways, right? But AI can actually create the granularity that you need, which will help develop and deliver highly targeted marketing and communication strategies which ultimately means that you can now personalize the content. Right? You can tailor the website content, you can tailor the email marketing, you can make the app interfaces. Like your app interface could look different from my app interface and that’s, I think, the goal of personalization, and how you can use AI to do personalization at scale.
David Arturi: Thank you, Dr. Perkins Munn, for joining me today and sharing your very valuableinsights. Look forward to catching up with you soon, hopefully, we talk in the near future and thank you to everyone who listened to our discussion today. To learn more about Lydonia, please visit Lydonia.ai. And for more interviews like this one, please visit CDOMagazine.tech.Thank you all so much, and we’ll see you soon.
