What Five Years of AI Have Taught Ally Bank

Penny Crossman: Welcome to the American Banker podcast. I’m your host, Penny Crossman. Ally Bank was one of the first to offer a virtual assistant with built-in voice interaction in its mobile app, back in 2015. Its vendor partner all that time has been Personetics. The two companies are now using artificial intelligence to offer insights to customers proactively. Here with us today are Anand Talwar, Deposits and Consumer Strategy Executive at Ally Bank and Jody Bhagat, President of the Americas at Personetics. Welcome Anand and Jody.


Jody Bhagat (President of the Americas for Personetics): Thank you.


Penny: Can you update us on Ally Assist. How’s that going? What kinds of

things are people using that for? And, what have you learned from rolling out this virtual assistant and being one of the first to do so?


Anand Talwar (Deposits and Consumer Strategy Executive, Ally Bank): Penny, thanks for having me. You know, rolling out Ally Assist was really interesting. As you mentioned, we rolled it out back in 2015, before the internet of things had really taken off, before the prevalence of voice interaction and digital assistants. So, it was interesting to see consumers get used to a new medium, as mobile banking was rapidly gaining traction and adoption. Quite honestly, what we saw way back when we launched was that people weren’t totally ready for the digital assistants that we were providing. Like I said, it hadn’t been prevalent. Siri wasn’t as popular, Alexa wasn’t out there. And so, because there was a lack of understanding as to what it was, we didn’t initially see the types of adoption that you would think that something like that would garner today, when voice assistants are commonplace in so many other mediums outside of banking.


It was an interesting learning experience for us in terms of deploying something a bit ahead of the curve in a space of financial services and banking, that hadn’t yet become part of consumers’ daily lives. So, great learnings in terms of how people interacted with it, what their questions were, really understanding the user experience and journey that someone goes through, and their readiness to adopt new technology, when it’s something that hadn’t been there. That’s guided a little bit of how we’ve approached the way we go about deploying new projects and how we conceive them right now.


Penny: Sure. So, I know in your relationship with Personetics you’ve evolved the technology that you’re working on together. What are some of the things you’ve added in the past five years, and especially recently?


Anand: Most recently, we’re really excited that we were able to build upon the digital assistant. The first thing that we started adding in concert with our partners at Personetics was an expanded library of personalized insights that we could bring to consumers – everything from a little bit of perspective into their transactional behavior… maybe their balance might be trending low based on the velocity of their transactions, or potentially scheduled payments coming up. So, a little bit of, “hey watch out for this.” Or, when we see out character behavior and just to make sure people are aware and staying on track.


More recently, that catalog has been expanding to include some more interesting things like educating people about savings and providing consumers with tips on how to start saving if we see that through their behavior that there’s some surplus and they have an opportunity to change their behavior in a positive way. It’s really fun…


One of the things we really try and do is create an emotional connection with our consumers. And part of that is giving them positive affirmation when they are doing really positive things with their finances. And so, there’s a fun insight that we provide called, ‘Savings Superstar,’ which provides positive feedback for a consumer who’s increased their savings month over month. There are a bunch of other interesting ones that really go from the transactional to aspirational. And that’s really helping to cement that kind of relationship that we want to have with our customers. And Personetics is a great partner in helping to deliver those things.


Penny: Do the alerts ever take a negative tone? Like, you haven’t met your savings goal this month, or you’re not on track, or something like that?


Anand: Not typically a negative tone, but we do try and make sure people are aware of their behavior. So, if they have done something like over-drafted or something like that, we’ll make sure that they do have those insights provided as well. But, we always try to make sure that they are presented in a way in which they’re positive and pushing people forward and not trying to pull them backwards or make people feel bad about their behavior.


Penny: Okay. And, Jody, can you tell us a little bit about how your technology works? For instance, in these examples of helping people improve their behavior, what kinds of data is the software analyzing and how does it produce insights?


Jody: Sure. And, just to build on what Anand talked about, we know that consumers are busy with their daily lives, taking care of their families and their work. And, we believe that innovative banks like Ally are in the position to really help customers anticipate issues and improve their day-to-day banking. That’s behind some of the more proactive insights that are being delivered now through the digital experience of Ally.


The way the solution works is that our algorithms analyze customer transaction activity. So, we’ll typically look at the historical transaction activity of the customer. And, at the point of login, all of the processing is done in real-time and on-the-fly. Customers log into the mobile app and Ally makes a call to Personetics. The algorithms then kick in to do the analysis of the transaction activity, and then responds back to the Ally mobile app regarding the insights that are most relevant for the customer at that particular point in time. It could be a duplicate

transaction, or the balance could be trending low, or is at risk of being negative, they could have an opportunity to save because they have a liquidity surplus in their checking account.

It’s the variety of these kinds of insights that tell customers that Ally is really looking out for them and trying to help them improve their daily banking.


Penny: The point about real-time is interesting because I know that sometimes with automated savings or suggestions of this nature, if there’s a transaction that’s still pending, that can trip people up because they think they have money in their account that they don’t have, or they have a different amount of money in their account than they actually have. How do you ensure that everything is real-time?


Jody: There’s an algorithm that looks at bounce forecasting, as an example. So, we look at not just the current balance, but we look at the patterns of activity, including anything that we can find around schedules and activity to get a better sense of what kinds of cash flow needs an individual customer may have moving forward. And, that allows us to then give better guidance in terms of when there may be an issue and help customers get themselves in a better position. That’s just an example of one of the insights and one of the algorithms. Other algorithms, such as looking at recurring activity, is less real-time and is more focused on looking at the historical data that identifies what are the typical deposit activities and payment activities that a customer has. So, when we see an anomaly that may be relevant to the customer, that will signal it for them and identify that this may be something that they should pay attention to. So, it looks at a

combination of real-time data and also historical data that we’ve seen over the past four to six months. And again, all that is done in real time, as the customer is logging in.


Penny: I know you guys have created something called, ‘Surprise Savings.’ Anand, can you tell us what that is and how it works?


Anand: Yes. We just launched a new set of what we call smart savings tools, on top of our online savings account. We did a lot of research with customers to understand what their needs were, what their emotional connections to savings were, what some of the barriers were to get people to save. And, after we spoke with thousands of people it came down to a couple of things that that we thought we could do to help people. And really, there were two. One was how people organize their savings, and help people optimize and maximize their savings.


To organize their savings we developed a capability, something like digital envelopes, to help segment someone’s savings on top of their account to help reduce the mental math that people do to understand what they have allocated to the things that are most important to them. We found that when someone names what they’re saving for, that emotional connection is the trigger to get them onto a plan and propel them to push forward. When it comes to optimizing or maximizing their savings, we developed things that we like to call “boosters.” And, ‘Surprise Savings’ is the featured booster and the one that is heavily powered by our friends at Personetics.


It leverages the algorithms that Personetics developed and evaluates the transactional behavior of a consumer… their transactional account, whether it’s with Ally or an external account at another financial institution. So, you don’t have to have your checking account with us to be able to use this capability. It analyzes their transactional behavior, what we expect in incoming transactions, scheduled payments, and develops what we call a ‘safe to save’ number. We then put some additional protections in place to make sure that number is truly safe. We’ll do a balance lookup to make sure there’s adequate money in the account. And then, we’ll automatically move that money from their transactional account to a savings account. We’ll do that a few times a week depending on if there is surplus available. It’s really for people who think they may have money, but they don’t know how to do the math to figure that out.


Leveraging AI and technology to help figure out the ‘safe to save’ number puts people on the beginning of an automated plan to help build their savings over time. And, we found that automation is that second key to create the snowball effect. That first step is getting started and helping people make that emotional connection does that. And then giving them a way to automate and keep the ball rolling was the other critical aspect. Capabilities like AI and an algorithm that we partnered with Personetics on is a huge part of delivering that through ‘Surprise Savings.’


Penny: And, is this something that your customers can turn on? Or, is it automatic?


Anand: They choose to turn it on and, if they want to turn it off, they certainly can.


Penny: Okay. So, overall, you mentioned in the beginning that you learned some things by people’s preferences, but is there anything else the two of you have learned from the five years of collaboration?


Anand: Having a long-standing partner that has a similar philosophy around customer centricity and innovation is critical. I mentioned earlier that the latest innovations and ‘Surprise Savings,’ we were able to bring to market very, very quickly because, we already had a successful implementation with our partners at Personetics. So, as they were innovating on things that helped us accelerate that, it took some burden off of the heavy lifting that we had to develop on our end.


When we figured out what customers really want and need in terms of saving solutions, it was kismet in terms of being able to work with an existing partner that we had some of the technology already implemented in our stack to help bring that to market sooner.


On the other side, there’s the customer learnings. Some of those things about exactly what types of insights to deliver to customers. I mentioned that we try and do things like provide consumers with affirming messages versus anything negative. It’s not that people don’t need the hard news sometimes, but in trying to help people really improve their financial lives, we found it much more powerful to be encouraging versus not. So understanding interaction patterns, as well as where to deliver those insights and when to deliver. You mentioned ‘real-time’ earlier. That’s critical to make sure we deliver the right message to the right person at the right time to help create the right behavioral change in people.


Penny: How about you, Jody? Is there anything that you feel you or your technology platform has gained through this partnership?


Jody: Absolutely. You know, we really enjoy working with leaders like Ally. And, as Anand mentioned, it’s sharing a common approach and a set of values around doing what’s right by the customer, leading in the marketplace. And then, going through continuous learning and optimization, as well, are both fundamental. I think what’s really exciting is this advancement in the world of automated finance and Ally leading with ‘Surprise Savings.’ I think we will continue to see rapid innovation in this space. We will see customers embrace these kinds of solutions when they trust the brand and they trust the algorithm. And, I give real credit to Ally for innovating in this space. I think it’s an exciting development for the industry as a whole.


Penny: Anand and Jody, thanks so much for joining us today.


Anand: Thank you.


Jody: Thank you, Penny.