Prospera Credit Union's Journey from IVR Phone Banking to AI-Powered Member Service Excellence

Welcome. So the title of the webinar today is, Prospera’s Credit Union Journey from IVR phone banking to AI powered member service excellence.

In this webinar, we will talk to Todd Teal from Prospera, and we will understand, the his AI strategy, how they’re using AI to do and part of the strategy.

And then, how do they kind of go about making a decision to go from IVF phone banking to the voice AI? And, Thomas shared his insight, on his journey going from the traditional IVF kind of technology to the AI voice.

So here are the presenters for this webinar.

Todd Thiel, chief digital and technology officer.

He will be interviewed by John. He’s a VP of customer success at Interface.

And during this, discussion, we will learn about, Prospera’s journey from, the IVR to AI voice, and have a broader discussion about, how AI is playing a role in the business strategy. And also the metrics they have seen, the impact they have seen, the lessons learned, the positives and negatives of, this kind of a project.

Then we are lucky to have Cortez Adam in this, fine summer dress.

And he will be helping us, with couple of demos of our exit pollutions and what is coming in the future. And again, with Jack Chawla, and I’ll be the host for this meeting.

So here’s the agenda. So, for people who are not familiar with about Interface, we will I’ll spend some time introducing you to Interface dot ai. Then we’ll go into the heart of this webinar, which is a case study with Todd, with John leading the discussion.

Then I’ll come back and talk about, evolution of AI. Obviously, AI is probably changes are happening every week. So we’ll try to simplify where the AI is going. And, kind of, spoiler that AI is going toward agentic AI.

And, we will show you the demo of agentic AI as it applies to our industry. So you get insight into how AI can actually help. Agent AI can help go beyond what you are traditionally seeing from, us and other vendors now now. So you’ll know where the future, is going.

And then finally, we’ll open it up for q and a. And in the q and a, questions, please feel free to kind of ask your questions as we are going along. And there’s a q and a area at the and you will see in the UI at the bottom, and you can start typing in the questions. Some of the questions we’ll answer as we kind of go along.

And some of the questions which are very interesting and more complex, we’ll answer them at the end of the webinar. The webinar itself is, ninety minutes. And I will strongly recommend that, you stick around and see the HDK AI demo. It’ll be very compelling to see where the future of AI as it applies to our industry is going.

Okay. So before I get started, a quiz, if you guys don’t mind, take the quiz.

Laura, if you can start the quiz, and let’s get the answers in so we could get a feel for, where you guys are in your AI journey.

Okay. And the answers are coming in. Have you currently deployed VoiceAI?

Looks like seventeen percent have deployed VoiceAI. So, people who aren’t deployed, you will get a lot of information in this webinar.

And it looks like a lot of people wanna deploy it at this year or, next year, which is good.

We’ll be able to help you. Have you currently deployed Chet AI?

Chet AI seems to be adopted much more. So I would say double our voice AI about thirty one percent have deployed to AI. And the people who are deployed, they do plan to deploy it this year, next year. So it’s fairly urgent for them.

Employ AI, about twenty one percent have deployed Employ AI.

And again, the plan is to to deploy it fairly soon. Some people actually a lot of people have no plans to deploy employee AI. Okay. So that’s where we stand as an industry. I think this is fairly typical of all the webinars I have done.

So hopefully you’ll get a lot of insights from Todd to figure out how what you wanna do around these projects.

Okay so let’s move on.

So about Interface dot ai, we have been in business for about ten years. We are headquartered in California. We have about one hundred and seventy five employees.

And we are one of the few vendors who are focused on only AI for credit unions and community banks, which means that majority all our people are focused on, obviously, this industry and majority of people are actually in r and d. So we are kind of, leveraging the AI and building compelling solution for our industry.

And this kind of, is different than a lot of players who kind of do what I call as a chat box AI.

So we are deeply engrained into AI, and we are doing deep work. And you’ll see where you’re going, with agentic AI, and you’ll find what we’re doing is very compelling.

Our effort have been recognized, by the the kind of the analyst and the industry press.

We have we have one of the few which have been featured in Gartner.

We have won many awards, Finovate and American Banker.

We have about hundred credit unions and community banks using us. You can see some of the logos here. Dupaco, EFCU, VioKey, and so on. And chances are there’s a credit union in your state who’s already using us. And chances are, they are using it at core you’re using. Again, online banking solution you’re using is the, phone banking, phone ACD or CCaaS solution you are using, which basically means that the work of integration has more or less been done. So, where if you decide to kind of use our solution, you will find that we can easily integrate with your, core and other technology stack.

Then from a solution perspective, we have taken this platform approach.

So we have invested ten years into kind of building out this platform.

We call it AI banking platform.

And this platform is where our one brain is. And this one brain is learning from all the data, all the interactions that we are having.

We have about one point five million conversation daily. Right? So, AI is a learning software unlike SAS. And every interaction we have, we end up learning from it. And that learning is built in into our platform and is developed across all our solution.

And of course, we keep learning as we kinda go along. So right now, there are three solutions in the market broadly speaking. So the first one is AI voice agent for call center. And the function of the AI voice agent is to help you automate the, incoming phone calls from, being the primary channel by which, the customers interact with you. And of course, in addition to automating these calls, there are additional advantage. Obviously AI is working twenty five seven, and it’s always learning, and it really can help you both from the member experience perspective and from the operating cost.

Then we offer, basically a chat AI, which is the way the way to think about it is like a AI copilot for your website and mobile apps. So when somebody comes to your website or a mobile app, you have this trusted, kind of a chat AI sitting there, chatbot or if you wanna call it, or a virtual agent. And it can answer any question they might have. For example, what is a perfect credit card for me? I’m a student.

Or, I want to, figure out how to reduce the, the credit card debt and so on. All the questions people have in the middle of the night or in on the weekends can, and can be answered by AI. And of course, both voice AI and chat AI can escalate to a human agent, whenever human agents are available. Right? So think of this as your frontline, kind of quote unquote virtual employees helping your customers twenty by seven with, sixty to seventy percent of the interactions which are coming in. And the rest can be escalated to the human agent.

So now as you can appreciate if, AI is handling the simple question, the comp the answers which go to the to the humans will become more complex. So, you probably want to figure out how to enable them to answer the complex question. So this is where the AI Copilot for frontline employees comes in. And it basically sits, side by side with your front frontline employees and helps them answer complex questions. Instead of escalating to the manager or saying, hey, I need to get back to you later, you can enable your frontline employees to answer the question immediately.

And the frontline employees can be either in the branches, or of course it can they can be in the contact center as well. So these are our three kind of the major solution, footprint right now. And then we are working toward Orbit, Spear, and Nexus, which are built on the agentic air. And we’ll talk more about this as we kind of go along.

Recently we entered the demo challenge at, American Banker.

And you recognize this guy who’s actually one of our panelist. And he actually did the demo of Agentic AI. And, Cortez will do the same demo, at the end of the, webinar. He had only six minutes at the American Banker.

And now he will have expended time to fully do the demo, and you’ll find it very compelling. So I’m looking forward to that. If you wanna learn more about that, you can kind of scan the QR code and see the press release. So we are very, very excited about winning this against the competition, which kind of validates, the vision we have for where the industry should go with the AI.

Okay. So let’s kind of put our solution of kind of a simple framework. So as I kind of talked about, we have this banking platform.

This banking platform has many many integrations with the different codes and so on. So all their work has been done as I mentioned before.

And within this banking platform, we have also built the biometric authentication. So that makes it very easy, frictionless to authenticate the people who are coming in through AI or into the contact center. And then we are also in invested in the caller ID forensics to help you reduce the fraud. Right?

So any call which comes in, we are able to analyze it across various different, points and tell you with, with a huge amount of confidence that the caller is who they are saying, they are. So we can detect whether so if there is a SIM swap, if there’s a new number, so we can raise a red flags for the calls. So you can pay special attention to those calls. So if you don’t see the red flags, you you can feel confident that the caller is, who they are saying they are.

Right? So this helps it is built into the platform. With other solutions, generally you have to go and acquire this as a separate solution, which is built into a platform. And these are the solutions on top of it, which we talked about.

And these are the solutions we are working toward, Sphere, Nexus, and Orbit, which are agentic. So basically, going on the journey from generate, from kind of the older technology of, convolution AI to generative AI to agentic AI. Okay. So let’s see one of those product.

I wish we had time to see show you all three. We don’t. So we’re gonna do a good good demo of voice AI only. So, Cortez, if you can unmute yourself and do a quick demo, please.

Alright. Awesome. Awesome. Thanks, Jack. Alright. Let’s jump right into an AIPB demo. Alright. First thing I’m gonna do is just gonna grab my iPhone here, and I’m just gonna give the bank a call.

Sure. Interfaces digital assistant. Please wait while I read the instructions. This call may be recorded for quality assurance.

For best experience, avoid using speaker mode.

How may I help you today?

I’m out driving around. I need to make a deposit. Are the branches open?

Yes. We are open. Our Main Street branch only has drive through capability today.

Do you want me to send you driving directions to the nearest branch?

Yes.

Great. Please provide me the ZIP code or city name.

Three zero one eight seven.

Thanks. There are a few branches that are open in your location.

The branches are open from nine AM to five PM, Monday to Friday, and ten AM to two PM on Saturday.

I just sent you a text with driving directions, and you can always talk to me whenever you need help.

I am available twenty four by seven. Is there anything else I can help you with?

Yeah. I just started a new business. It’s not going well.

I may not make payroll. Can you help me skip a loan payment?

I can surely help you with skipping a payment. I’ve sent you a text with the link to the application to skip a payment along with the instructions.

Please click on the link and follow the instructions.

Please note that the application should be filled at least five days before the payment due date.

To assist small businesses like yours, we’ve removed the thirty five dollar fee associated with skipping a payment. Do you wanna know which loans are eligible to skip a payment?

No. Can you tell me the balance in my checking account?

Sure. We’ll help you with your account balance.

I see that you already have registered for device biometric.

Would you like to verify using your device?

Yes.

Great. I’ve sent you a link to your phone. Click on it so I can verify your biometrics.

Thank you. I verified your device.

The balance in your checking account ending with two two three four is two thousand three hundred and thirty seven dollars and fifty five cents.

Also, I wanna remind you that if you maintain a minimum balance of three thousand dollars, you will be eligible for an additional one percent cashback.

Do you wanna know the balance of your savings account?

Oh, I’m traveling, to London soon. Can you help me put a travel notice on my account I mean, on my card?

I can help you set a travel notice. I’ve noted that you’re traveling, and I have successfully set a travel notice on your card. We’ll make sure your transactions run smoothly. By the way, you’re traveling a lot, and I recommend you switch to our new credit card with no foreign transaction fee and many more benefits. Do you wanna know more?

No. But my wife is looking to buy a car. Can you help me with car insurance?

Sure. Happy to do that. Give me a moment. I will transfer the call to our specialist, John.

Awesome.

Well, guys, hopefully, that was helpful. Just a quick walk through. Five use cases that really highlight the functionality, the depth of integrations, and the ease of use of our AIPB product. Our voice AI tool is meant to come in a natural voice, but can be customized into my voice or that special employee at the credit union or at the bank who everyone loves, or maybe even that CEO that’s getting ready to retire or that board member that everyone just loves talking to.

We’ve got that functionality kind of natively built in with that custom voice option.

So with better than human understanding, tight integrations across the industry, allowing not only for dynamic informational knowledge to be shared with the users that are calling in through the voice channel, handling automation just like Jack said, but we’ve also got that ability to just dive deeper and deeper with those third party integrations. So, Jack, I’ll turn it over to you to get into some of the high level capabilities further.

Sure. Cortez, you stole my thunder in terms of the capabilities, but thanks for showing the demo.

And as Cortes mentioned, I mean, what you saw cannot be done by IVR.

And it was almost like a fairly senior, human agent kind of experience, with this AI voice technology and talk talk more about it as we kind of go along. So as, Cord has mentioned, the voice you heard was very human like. Right? And you can do custom voice.

So if you have a, you are working with a personality, celebrity with with your brand, you can use their custom voice, for example. Or as Curtis mentioned, you can have your, CEO or a special employee’s voice, as greeting the customers and talking to the voice. So basically what happens is, we can record the voice sample of that employee or the celebrity you are associating with, and then we can convert that into the voice of the AI voice. Okay?

And as Cortez was kinda talking, he was actually on the speakerphone. Even though we don’t recommend speakerphone, the AI was able to understand all, Cortez was saying and able to respond to that. Right?

Of course, the quality will the understanding will better with without speaker phone. But the reality is people will talk while they are in the car and so on. And you can see that it can actually understand it very very well.

Then you saw, Cortes holding up his phone, and doing kind of the face ID, which is the device biometric function. So depending on the phone, you can do the biometric using your face ID or the fingerprint.

And as you can see, this is like, frictionless, from the customer member perspective.

And from the credit union community bank perspective, it’s very very secure. Right? You cannot fake, biometric unless you are in a special Tom Cruise kind of a movie. It’s really tough to do this in real life.

Then the caller ID forensics, we talked about that. Unfortunately, we cannot show you all that concepts. But basically, we are listening to the call and, looking at the footprint of the call from various different angles and making sure that the caller is who they are saying they are. Then, unfortunately, AI cannot handle hundred percent of the calls.

So that has to be transferred to the human agent. For example, if somebody wants to do something complex like, open their trust and so on or want the quote for the insurance, we do the transfer. And the transfer is very contextual. So it will take the transcript.

So Quartus and the AI talking that was transcribed, and it is provided to the agent. So they know exactly what was talk, what was the discussion. So the agent doesn’t have to start from zero. They can see what the transcript was.

It will be summarized. You will have the sentiment. And then as a human agent, you can continue from there. So you start starting from scratch.

Our technology allows you to do complex interactions as well. So you can do, kind of a joint account, interest account. When you’re evaluating AI, you will see that majority of the solution out there are not able to handle this.

Then if you paid attention to Quartus discussion, you might have noticed AI prompting Quartus on different things. For example, if you maintain a certain balance, you get additional cash back. So this is the upsell and cross sell motions.

So depending upon your business, we can kind of work with you to build those suggestions into the conversation. So it comes across naturally in the context of what the customer is asking, and help you position additional products and services to the caller.

Then the other things, insightful in analytics.

So as I mentioned before, AI is a learning software like SAS. And AI is able to obviously answer lots of questions. But the questions it doesn’t answer, it transfers to the human agent. And that insights are available to you. So you can kind of go to the analytics and say, okay, what are the things which AI is not able to handle? And what is the frequency and so on and so forth?

Where is the person getting frustrated? So all the analytics can help you train the AI to get better every day. So, depending upon your kind of the effort you put in, you can make AI pretty much better every day. And then you can kind of deliver more, kind of benefits to the member, more automations, from your perspective.

So this is very important kind of a concept Like, you bring the AI into the world, it will know certain things. With the analytics, you can make it smarter every day.

Like, as in real life, like, once I understand something, I can simply move on. So if the member or the caller says something, if you understand something, you can do the interruption in real time. You don’t have to wait for the AI to finish the sentence. So that makes it more natural.

Then when the human caller is talking, we are actually transcribing it in real time. So we are able to answer, immediately when the caller stops talking rather than waiting for few seconds, which is a kind of dead time. So you will see in some AIs, you will see, like a noise being made before the AI can answer. In our case, we are able to answer mostly immediately.

And then you can leverage the text response. So for example, somebody asking for driving directions, to the to the branch, you can just text the link rather than having to kind of give it over the phone.

So this is the AI voice capabilities.

Now, we didn’t show you the demo of AI chat, but brief let me briefly kinda talk about what AI chat can do. So AI chat is basically a copilot on your website and the mobile app. And the first thing it does is checkgripetize your website. So just like checkgripet is sitting on top of Internet and giving you the perfect answer in most cases, to the question rather than giving you ten links. Same same thing we do, but with your website and with with your content. So it is highly grounded into the information you want to learn from, which also means that you don’t have to sit down and figure out, okay, here’s a hundred and in intents and here are my hundred answers.

You simply say, this is my website and this is the content I want my, AI check to learn from. We’ll automatically learn from it and start answering those questions. So you can get started within a day or two to basically be able to answer any questions, people have coming to your website. But we don’t stop there. We do phone banking kind of a concept, on the website as well. So that what Kotas did on the phone, that can be done, through the, through the through the AI chat itself.

Obviously, people can do that through online banking, but sometime you wanna simply do it through conversation.

So you can go from searching for something to doing transaction. It happens all in the same conversation, with the visitor on your website.

Then the other cool thing we have is smart conversion.

So as part of the journey on your website, people might want to apply for credit card or a loan and so on. Generally, they have to fill out a form. So we are kind of helping with that as well. So when people are moving from field to field in the form, we can prompt them with information.

For example, somebody says, the VIN number. You’re asking for the VIN number in the form. People might not understand what the VIN number is. So AI can prompt them saying, hey.

This is what the VIN number is. This is where you can find it. So it is basically hand holding as you’re filling out the form. And we have seen that, with the smart conversion, you can double the conversion rate for, your form.

So which means a lot of new business for the credit union and the community banks.

Device biometric concept applies to AI chat as well. So, obviously, if you’re doing phone banking kind of a concept on chat, you need to authenticate the user, and you can do that through device biometric. Same option and cross sell concept apply to that, AI chat as well.

And then, you can escalate to phone. Like, a lot of time people think that to use AI chat, you need live chat as well.

But a lot of our customers actually don’t do that. They simply have AI chat on the website. And if the escalation needs to happen, they do it to the phone. The reason is the AI chat is smart enough to handle simple questions, on the on the chat side. And, if the escalation needs to happen, the questions are fairly complex.

They are generally not easy to handle through the live chat itself as well. So it’s better to simply go to the phone rather than have a live chat. And of course, it simplifies your contact center as well. So you don’t have to have different types of agents doing phone chat and and, and, phone calls and live chat.

You don’t have to train them. As I like to say, like some people do video as well. So I as I like to say, it’s very hard to find an agent which is good looking, speaks well, and types well. Right?

So you don’t have to go to that and kind of try to optimize, your your agents for that. You can train your agent to do phone and let AI do the rest. Right? So think think of like, instead of going omnichannel, you can go omnichannel, and then you focus your human resources on the phone channel, which is the most effective to handle the complex issues.

Then multimodal UX, we can answer the questions using the carousels and so on. If you do use live chat, we can integrate with live chat. And then we provide insightful in analytics similar to the phone system.

Then the frontline employee, a lot of people might not be familiar with this concept, but think about, like chat GPT for your internal policies and procedures.

So, this frontline, employee AI, you will upload the documents or point them to your share SharePoint, and you’ll learn everything about your policies and procedures.

And then it will be able to answer those questions, for the agents or the branch employees.

You type it in, you get the answer. It is fully grounded into your document. And when the answer is given, it will point to where it is coming from. So that’s kind of one concept.

Obviously, if you’re transitioning from the phone, it will bring the conversion summary and so on and so forth. And then we also provide automated actions, which is basically, automating like a, RPA, robotic process automation. All the things which agents need to do again and again. For example, block log the card, or, change the address and so on.

So this is all done to the automated action. And then, we have, internal built in knowledge management and role based access control. So that’s kind of a, high level overview of our solution. We’ll be happy to engage with you and show you the customized demo.

When we do the demo, we can point it to your website and show all the things kinda we talked about.

Okay. So now let me turn it over to John. So John, if you can unmute yourself and Todd, if you can unmute yourself. So John will take it from here and talk to Todd about their experience.

Great. Jack, thank you so much. Thanks everyone for being here today. Really excited to be here with you and to be speaking with Todd Theel, the chief digital and technology officer at Prospera Credit Union.

So, Todd, we’d love, to learn a little bit more about you and the credit union.

Sure. Thanks for having me, John. It’s great to be here. Yeah. So I, lead technology, everything digital here, at Prospera. I’ve been with the Credit Union three plus years.

And so, yeah, just been leading the charge in terms of where we’re going from a capability perspective, with a lot of, support from a lot of great people here. So we’re excited to have this opportunity today.

Great.

So I think that one of the first things we’d love to learn a little bit more about, Todd, is, you know, how is Prospera’s business strategy delivered more effectively with AI? What is it doing for you?

Yeah. Absolutely. You may have gathered from our wallpaper. Our our mission is pretty simple. We’re here to help you prosper, John.

That’s where we focus our efforts early, define what that means for you, and it’s prospering financially to enhance your quality of life. And when we talk about AI here as an organization, we really talk about it in terms of business strategy first. We don’t really talk about here’s a, you know, kind of quote unquote AI strategy. It’s really about the business strategy and how are we solving problems or taking advantage of opportunities with AI.

And when we look at our business strategy, pillars of that are are what you see here on these slides. So there’s this member centric piece, a team member piece, which is really our employees, and a piece focused on risk. And when we look at where we can drive, you know, improvements in those things on the member centric side, we’re looking for those personalized solutions, you know, like you heard in the demo with Cortez today. Right?

Some, AI that understands something about me and can add value, to me and my financial situation in some way, shape, or form, and really have that tailored advice and offer self-service. We see that coming more and more, where people obviously wanna self serve, to a fair degree.

On the team member side, it is all about efficiency, making things easier for our staff to to to help our members prosper. So we have, different groups of categories of pain points and things, from a staff perspective. How do we how do we make their jobs easier, so they can spend more time, you know, with the members helping on what matters most. And then in terms of risk, you know, of course, you know, fraud is a huge challenge for everybody these days.

Cybersecurity and so forth. And so there’s a lot of good AI solutions out there. And, Jack kinda touched on a few of those as well, you know, with the caller ID forensics and some of those capabilities too. So we’re looking at solutions, that can solve problems or take advantage of opportunities, any one of these three pillars.

Great. Thanks so much, Todd.

One of the great things about the relationship with Prospera is that we’ve been able to, you know, roll out our AI voice product, which has been live since twenty twenty three, and also really support Prospera’s business through integrations with Scimitar and RingCentral. You know, one of our focuses at Interface is creating a a connected ecosystem that is able to really easily deliver on those transformation goals.

So glad to see that we’ve been able to partner it and build this out for you.

Absolutely. Yeah. You bet it’s been a great partnership, and we’ve had a lot of, success and learnings along the way too, of course, that I’m happy to share and talk more about here.

Awesome. So I know that we also have, we have a a sample number here for for folks to call.

You can save it until after the webinar. We wanna make sure you’re able to kind of get all of Todd’s great insights. But after the webinar ends, we’d love to, to ask folks to take a call in, experience Prospera’s VoiceAI, really see what this is doing for their for their members and for, for really driving forward their innovation agenda.

So with that in mind, you know, Todd, as we’re talking about, you know, business strategy as a primary focus, but also AI strategy as being a part of that, what are some of the problems that you were really trying to solve with Voice AI?

Yeah. Great question. We were at a crossroads because we were, going through a core conversion, like a lot of people on this call either have done probably recently or maybe they have one on their radar. And so part of that core conversion, we we had on our old core system, a traditional, you know, IVR type telebanking solution.

We were gonna get a new solution when we moved to Semitar with Jack Henry. But we had to decide if that was really the kind of the strategic solution going forward, or is this if we’re gonna go through that core conversion, is this the time to move to an AI type solution? So we had that fork in the road to consider. And as we were looking at that, then we were, you know, kind of investigating and talking to a lot of the peers about what to expect with go live and call volumes, in terms of increases.

And we were hearing, you know, fairly consistently, you know, expect, like, a three x to five x kind of increase in call volumes.

So that was weighing on us as well. And how do we scale to handle that? You know, we’re we’re a midsize credit union with a small call center, maybe five, six, you know, permanent staff members in the call center. And so, you know, if we’re gonna increase volumes that much, what do we do? Do we temporarily add staff? Do we look at other solutions?

And so we saw a unique opportunity to look at something like Interface and think about how it could solve those problems and also really, just set up the foundation for the future.

Because if we can have something that integrates with a lot, you know, not just the core, but also our voice over IP system with RingCentral and some of the other solutions and vendor partners that we have, then it really does set us up, you know, strategically, to continue to build on these capabilities. And it’s not just a a a one specific point solution.

Great. And and how did the employees respond to, you know, some of these these new innovations?

Yeah. I think for the most part, the staff was pretty excited. You know, that this is back in, you know, twenty two. We started researching at twenty three.

So AI was this, you know, kind of big newer thing for a lot of us, and understanding what that means, and seeing where there was, you know, capability and possibility to provide a much improved, member experience. Again, kind of, you know, just going back to our business strategy and the focus on the member, you know, what does that look like? How is this gonna help them? And when we looked at our call center, we had challenges like a lot of traditional call centers, you know, in terms of answer rates and staffing and turnover and all those types of things.

And we saw some some potential for this to help, you know, alleviate some of that.

And and so, yeah. So overall, I mean, there was a lot of excitement, and we made sure, you know, one of the lessons we learned early on was to have, you know, a a coalition of the willing, you know, really kinda get involved early and roll up their sleeves, if you will, you know, a small group, you know, five ish, you know, type people, different areas of the business, you know, that were engaged early and often to put to come together, look at these types of solutions, talk about the problems we’re going to solve, and make sure that everybody, you know, felt comfortable and could really be a champion of that. This couldn’t be just a tech thing, couldn’t be, you know, just a a cool AI type thing. It had to solve a business problem or problems, and we had to have multiple people from multiple areas championing this to be successful.

Absolutely. So thinking about that, you know, that that choice to, you know, engage with your with your call center agents, roll something new out, take advantage of this new excitement around, all the evolutions in AI, how did you think about the partner selection process? You know, where did you look for those, and, ultimately what led you to Interface?

Yeah. I mean, it we started, you know, with a wide net like a lot of folks do. So, you know, going to, you know, different, you know, fintech forms like Ventures Fintech. I remember, you know, that at that time it was CUNA, but TruStage puts on.

Looking at a lot of webinars, talking to a lot of peers, you know, doing research, that type of thing to kinda come up with a basket of possible solutions to solve this problem. And then we started to apply filters from there. And one of the early filters we talked about is we want somebody with financial services, specifically, you know, credit union expertise, because we those that are in the credit union space, it’s it’s unique. Right?

We have a we are so member focused. We don’t have customers. We have members. It means different things.

And so we really wanted somebody who had had a proven track record of success, in that arena. And so we we applied multiple filters, you know, credit union focus, proven track record, experience integrating with our core and our VoIP system. So that just narrowed the list, you know, further, further, and further down, you know, fairly rapidly. So we ended up with a short list.

And then we really kinda rolled up sleeves and dug in, you know, on that short list, you know, that we were looking at. And in terms of why Interface, I can I can touch on that a bit more too? You know, as we started to work with Interface, you know, it was pretty obvious early on that they were really confident and comfortable, you know, with our core and and what we had there. We were impressed how quickly they got things up and running in a test environment.

Even from a demo perspective, they could demo stuff awfully fast. We’re like, okay. I mean, we were more worried about the complexity, I think, than they were because obviously we hadn’t done it before.

And so, you know, just doing the demos, it was it was obvious that they had had that experience. And in talking to, you know, the reference customers and peers, there was a credit union, in Wisconsin that was already a partner of Interface. They had a great experience with implementation, and it just kinda built on that confidence.

So that was one of the things. And another was, you know, specific to our voice over IP system with Ring. You know, this was a vendor that that a lot of at these times, a lot of these AI vendors had not yet had a partnership with or done any types of integration with, so it was new. And that’s a lesson too is to understand, you know, what level of existing integrations are there for the vendors that you have in house, which one of them, you know, would be brand new to any of these AI partners that you’re considering, and just expect the new ones to take quite a while.

Right? Because it’s because you’re plowing new ground. And even though, Interface had done some work with Ring, how we wanted to set it up turned out to be a bit different with Ring. So we ended up plowing some new ground too.

And it was no fault of Interface, no fault of Ring. We just figured out, hey, we want it to work a little bit differently. And so those become kind of long poles in the tent, if you will. So integrations are key.

Having people with that experience and kind of a proven track record of those integrations can can make things very smooth and save you a lot of time.

Awesome. So speaking a little bit about that, that discovery process and moving beyond, you know, moving beyond, the decision point into implementation. What did that journey look like for your team? What did you have to do and how did you experience the process?

Yeah, absolutely. So I mentioned we had a small group of champions. That group stayed on then, you know, throughout the entire implementation.

So a cross functional team, you know, everybody wearing, you know, kind of multiple hats. We’re not a huge org. And we also didn’t want a massive implementation team where it just becomes, you know, more cumbersome.

And so we basically had a dedicated project implementation team, from Interface.

And then we met, you know, at least once a week, but we always had had, you know, the option to meet more frequently. And Interface really guided us early on, and I alluded to it. I mean, they had a very standardized process for onboarding new clients, even back in twenty three. I’m sure it’s that much more mature even now, but I was impressed even then with, you know, how many standardized templates they had and how easy it was to say, you know, this is how we want these things to work. And so if you if you haven’t had a chance to go deep on these solutions, what you’ll find is that organizations like Interface have already predefined, experiences or they might call them intents.

And so they’ll come out of the box and say, hey. We’ve got x number of predefined intents that work with your core. That’s fantastic because then you’re not building new bridges to a to a new destination. You’re just leveraging things they’ve already built and it’s very fast for them to implement those things, which is wonderful.

So that we we went through all the standard intents and said, yes. We want this one. No. We’ll wait on this one.

Yes. We want this one. That part was very straightforward. And then they gave us an opportunity to do some custom intents and say, yeah.

Let you know, we want some unique experiences. We wanna have this handled this way because it’s, you know, prosper and this is how our members wanna interact, type thing. And we came up with a list of those.

And so that that was a a regular cadence. You know, we had a test environment running, you know, within, gosh, probably six weeks, where, you know, we had active things integrated.

And that and that includes the time for our core partners and these other partners to do their pieces too. Right? It’s not just work that has to happen on interfaces side. It’s these other partners too.

Yeah. And then from there, you know, we went on and then and then we came to a big decision, which was when you roll out these solutions again for those that are new to this, you really have, have kind of two main options when you start. And one is to do kind of an informational only type experience when you go live, and another is to include transactions with that experience on day one.

And so initially when we were thinking about it, we thought, well, you know, again, this is brand new and our and so we were timing this to go live with a core conversion. So we didn’t just have a new AI, you know, IVA. We had a new core. And at that same time, we had new online banking. So this was a very big bang.

And I’m not And I’m not all at once.

Right. Right. And I’m not sitting here recommending that that that’s for everybody. But given our situation and what we were coming out of, it made a lot of sense.

And we said, this is what we’ve gotta do. So we’ll we’ll do that. And then, you know, we had an honest kinda heart to heart conversation with Interface. I remember this like it was yesterday where we were saying, you know what?

We’re because of all these changes happening at go live, we’re we’re a little more, you know, we’re conservative or Midwest. We’re, you know, we’re a little trepidatious about trying to do both information and transaction at once. Right? So let let’s just stick with informational.

And what I mean by informational is, you know, what are your branch hours? Are they open or closed? You know, send me a link of direction, you know, that kind of stuff versus transactions and authenticated, you know, transactions to your core. And so so we were thinking at go live, we would just go with informational, and interface.

We would do a phase two and come back with transactions. Well, they they sat us down and said, look. We’re confident we can do transactions on at go live as well. And, you know, Todd, you’re gonna be better off because your traditional telebanking people can do transactions today on your old core.

Right? So rather than have it be a takeaway, what if we made that all happen and go live? And so they convinced us that they that they could do this, and they did it successfully. So they got transactions up and running very quickly.

We went through, did a a ton of testing in a short amount of time in amongst all the other core conversion and online banking testing, and it all went live on the same day. So information and transaction. Now again, not saying that’s for everybody, but from for our situation and where we were coming from, it it made the most sense and it and it works.

No. That’s great to hear. One of my favorite parts of my job in in customer success is obviously hearing that this whole process worked well and that you got what you wanted right out of the box.

Right. Right.

So, now that the the solution has been live for a period of time, you know, what kind of impact have you seen on the metrics? You know, how is adoption looking, and how are you thinking about that overall that overall journey?

Yeah. I’ll I’ll bifurcate this just a little bit. I’ll talk about what happened at conversion and then where we are, you know, eighteen months later, which kinda tells the big story. At at the conversion, basically, like I mentioned earlier, you know, we were told by peers, you know, expect three to five times call volume. We ended up with ten times call volume. So we were very, very thankful that we had your face up and running because our our typical, you know, two twenty calls a day became two thousand five hundred calls on a Monday, when we went live. And so when that happened, you know, one, you know, the IVA picked up every single call, you know, and two, you know, it basically handled like forty three percent of them on that very first day where they did not get transferred, you know, to the call center.

And so we had temporarily staff up the call center because, again, just trying to prepare, you know, hope for the best plan for the worst.

And so, you know, they they still took, you know, a large portion of calls, but nothing like they would have if we hadn’t had, you know, this the IVA in place.

So that that was significant.

And in the first five days, you know, we we go back and look at metrics. You know, we had a hundred and forty two hours of IVA time where we’re talking to members just in the first five days, you know, of this system. So that was three and a half people for eight hours a day for five days straight is what that math looks like. So it was pretty impactful for us, to have this in front of the organization and to have it available twenty four seven. Right? Because especially during a core conversion go live and online banking go live, all these changes all at once.

So so we felt really good about the initial, you know, use of the product and how it was received.

You know, kind of fast forward eighteen months later, and if I look at, you know, just before go live, before we had this in place and we were traditional, back in twenty three at this time, we averaged about fifty seven hundred calls a month. In twenty twenty five now, last month, we’re over eight thousand. So we’ve got, you know, call volume has gone up, you know, quite a bit.

Back then, you know, we would have an answer rate, you know, on certain days in in the sixties. You know, I mean, just, you know, with staff turnover, people are out sick. You know, the wait times have gone up, you know, these types of things.

And you look at it today and I’ve I’ve, you know, looked at these numbers very recently for last month. You know, we’ve got an answer rate in the nineties.

We’ve got a wait time of a of a minute. I know the data is a minute and seven seconds average wait time for our call center, which is so much better, than it used to be because, you know, we had and we haven’t had to increase the staff because, again, we’ve got the IVA taken a very healthy portion of these calls.

And so one of the things that we did see, and you may hear about this and expect this to a degree, is that our average handle time went up post go live, and that’s that’s logical because if you think about it, now our members, a lot of them are using the IVA for the simple things. Right? So they’re saying, I wanna check my balance. I wanna, you know, move some money between my accounts.

I want some loan information. I got a credit card, you know, issue, what have you.

And so that stuff gets automated. Now it’s the more complex stuff that comes back to your call center. So, our average handle time right now is four minutes twenty one seconds. We’re pre go live. It was three minutes forty three seconds. So just about a minute longer than it used to be. But I’d look at that as healthy because the complex stuff is getting to our people, and that’s where we want them spending their time.

Mhmm. Absolutely. Put people on places where people, you know, work the best, which is complex questions, building relationships, and answering those tough questions.

Absolutely.

Great. And what we see here on the screen is a little bit of information, about Prospera’s, automation trends as as Tod just mentioned. And what you see is, you know, since go live and as time has gone on, we’ve continued to see an improvement in automation, and automation being essentially the percentage of calls that are able to be handled by, by the AI and not being transferred to to an agent. And, you know, one of the things that we have worked with, Todd and his team and with all of our our CU partners, is essentially on on value engineering and understanding what are those intents that are live today, what are those additional intents that we can add to continue to improve automation?

And how can we continue to drive more and more value through AI, and re and relieve some of that burden on your call center as part of our our full solution and managed service offering? This is one of the one of the favorite things that we get to work on. So, Todd, I just spoke a little bit about, you know, our approach and on on value engineering and managed services. But as you thought about it, you know, internally from the CU, how did you think about increasing those automation rates over time?

Yeah. A couple different ways. One is, you know, talking to our call center manager and understanding, you know, what do we have for dispositions there? You know, what’s making it through? And from that list, what are those things that shouldn’t make it through? Because we can really turn it into an experience or an intent.

And then we kinda look at that data and pair that with the tools and the analytics that Interface provides. And Interface has some pretty robust, analytic tools, you know, that come as part of this package. You don’t have to pay extra for them. It’s all built in.

And you can see it on your screen here. But really being able to compare and look at, you know, the intents and that these are these experiences. So looking at the data, you can pull up these lists, and it’ll say, hey. You had x number of people ask about a loan payoff or order new checks or what have you, and that is already a supported experience.

Would you like to talk about enabling that?

And we’re like, sure. You know, of course. Right? So, you know, let’s put that in UAT.

Let’s test it out. See if we like the experience, need to tweak it at all, what have you. They just make that very easy, which which is fantastic. So we’re not trying to call through, you know, massive, you know, data stores and and run, you know, very sophisticated analytics ourselves to understand, you know, what’s available and what we should be focusing some of that attention on, which is great.

And another thing we do then was we tie it also to other things that are happening in the business. So, you know, let’s say we’ve got another system change coming out, or we wanna run a campaign, you know, a a refinance, you know, or something like that. You know, we we can leverage the system very easily, you know, working with our partners and interface, you know, to put in a new intent, a new experience, you know, a new, kind of brand campaign, if you will, to take advantage of what’s going on, you know, from a seasonal perspective.

And so it offers a lot of flexibility that way, and we we’re still finding new kind of creative ways to use that.

So here’s another question that I think will address some of the items that are coming up in the the q and a chat as well is, how did integration with Scimitar help, with a faster response and reduce your handle time? How did you work with that core system?

Yeah. And so I alluded to this, you know, the integrations are already built and proven with Semitar. So they they work and they work very well. And then one of the things we really liked about that integration is that when you transfer calls in to the call center, if you’ve done any kind of transaction in the IVA, it’s doing basically a screen pop, and it’s bringing out that member’s account in the core.

So no longer, you know, do I have to ask you once again for your member number? Do I have to, you know, as a call center rep, do I have to type it into the core and pull it up and make sure I’ve got the right account? All that stuff pops up. It’s seamless, and that’s right next to another screen that has all the chat content in it.

So I can, you know, that’s basically my broad over from the IVA. And so I can see, you know, what you’ve done already in the IVA. I can see, you know, where you left off, what questions you had, and then I’ve got this account that’s already open in my core. So that brings a lot of efficiency in and of itself.

And then the additional things we’re looking at are some of the products, you know, that that Jack alluded to, where, interface offers solutions to go, you know, look through your knowledge base and pair that with with what is coming over from your IVA. So where we’ll be going in time is being able to take those questions that were in the IVA that got transferred to the call center, automatically doing the knowledge based search, summarizing that information with interfaces AI, presenting that summarized information back to the rep. So now the rep isn’t having to go in. You know, I’ve gotta go ahead and, you know, how do I change a pin on a on a debit card?

And maybe, you know, maybe I don’t remember. And I’m going in my knowledge base and looking that stuff up. Well, now, you know, Interface can provide a solution that will summarize that, from a procedure that is a knowledge base and present it back to the rep. So those things are coming together.

It’s all kind of part of the strategy that Interface has, to to really make, again, that that team member’s life easier in serving the member.

Great. And as we, we get close to the end of the, of of of the the panel discussion here, love to learn just a little bit more about some best practices that you can share with your peers on the call, with other financial institutions who are looking to adopt AI solutions.

Sure. Sure. Yeah. And I’ll, I’ll recap a couple things too. So, you know, one is start with the, start with the business problem or the business opportunity you’re going after and, and get alignment on that.

We all have a list of of pain points as organizations, and we all have a list of of new things, you know, we wanna be doing, and make sure, you know, that your your senior team and your peers are all aligned on what that list looks like and what’s at the top, so that you aren’t wasting time. Right? If you’re in if you’re playing a key role in investigating these kind of things, you wanna make sure that you’ve got the support, behind that to say we wanna spend, you know, time and energy and money here, solving these kind of problems or taking advantage of this. So that’s the first thing.

Start with the business problem, you know, get alignment. And then like I had shared, you know, we we had a small coalition of the willing saying, yes, we wanna be a part of this particular thing. We wanna be on this exploratory team. We wanna look together, try it together, this type of thing.

I think that’s huge if you can get so that it’s not just a technical thing. It’s not just an IT thing. It’s a business thing. Right?

It’s a business project, to solve business problems or or create business opportunities. So try to form a small cross functional coalition, and bring everybody together. You know, once you’ve done that, you know, again, looking at partners, you know, in your industry, obviously, and and we’re here, you know, on a CU call. So whether it’s a CU or banking, make sure that partner has the expertise, that you’re looking for in that industry because every industry is different obviously in a lot of ways.

And then the integrations are big. So if you’ve got unique solutions, or things that not every vendor partner has, you definitely should put some weight on, that a certain partner has experience with a certain core or a certain island banking system or a certain VoIP system.

Just because it can save you an awful lot of time in implementation and dealing with, you know, first of a kind type problems.

You don’t have to plow any new ground that way.

Great. Thank you, Todd. Thank you, John. You kind of, heard about what Interfacial AI does.

From Todd, you heard about how they’re using our AI voice solution.

Now let’s talk about where AI is kind of going. Right?

So we kind of, before ChatGibri came along, all AI was conversational AI.

And, the application of this AI in the in the business world was limited and, not very successful, frankly.

So these are the experience you might have with, like, vendors like AT and T deploying, chat AI, and, most of the time frustrating the members, the customers.

But they had to do it because it’s a kind of a high volume, low value kind of business. And, they wanted to really reduce the number of phone calls coming in. So they tried everything they can to reduce it. But Cognizant AI was really not that successful, in achieving the goal. Especially for our industry where the relationship is very important.

The Cognition AI, if you’re using it or if you’re still using it right now, I would say get off it as soon as possible. Well, it doesn’t really work very well and frustrates the members.

So after ChatGPT, we are kind of entering the generative AI age, and it’s doing wonderfully well. So generative AI, the understanding of the natural language is amazingly good. Understanding the intent of the of the typer or the caller is amazingly good. And the answers which are coming back, obviously, they are, need to be grounded, but they can be very, very natural.

So, generative AI, is proving to be very useful. And this is where, we are, right now as a as a as a interface dot ai. So we are end to end generative AI across our technology stack. We are in the process of implementing it with all our customers.

And, our chat AI and employee AI is already generated AI. And the phone also is getting to the generated AI. So this will be, like, amazing experience. So I think we are going from conversation to generated AI. It’s like a ten x improvement.

Then, agentic AI, if you are using chat GPT and you’re doing deep reasoning, deep research and so on, you already have experience agentic AI, in the consumer world. I personally use it all the time.

And it saves me days and days of research.

It takes obviously long time, fifteen, twenty minutes to do the research, but it still saves me couple of days of research to get me the answers. So this is like deep reasoning and so on and so forth. Now how do you harness this agentic AI in the business world is where we are focused on. And this is where we’ll show you the demos of how we are leveraging agentic AI. Similar to what we did with general AI, make it grounded, make it useful in the business world. Same thing we are kinda on the journey of for doing with the agentic AI.

So in the past, with with the with the technology we had, we kind of lived in the omnichannel world, and we try to bring AI into the omnichannel world per channel. Right? And, we try to do as best as we can given the technology and given the, the approach.

But that is in the past. So now I think at least interface dot AI, we are able to give you a unified experience. A unified experience means that you can look at AI as a Omni AI, and you can apply to all the channels with one brain. And the channels can be either, you can continue using them all the channels you have, or you can simply say that, unified AI approach, the Omni AI approach is so good that you can handle all the simple interaction and take away some of the channels, from the live perspective.

And then, like the thesis is that if you need to go to the human, it is gonna be complex. Then use the channel which is the best for human to human interaction. And the AI let AI be the omni AI omni channel and handle the interaction across all the channels. So for example, somebody coming to website, AI check can handle it.

Somebody coming to texting you, AI check can handle it. Somebody coming through social media, AI check can handle it. And then whenever you need to escalate, you can probably escalate to the phone channel or a video channel, or an appointment in the branch.

So that’s where we are, with this unified approach.

And then we are kind of working toward something which is, kind of radically different in a way. So, we can kind of invent the future of autonomous AI. The AI which is kind of, not only doing what we are kind of used to, or at least have a good understanding of from, automation perspective.

You can think about agent AI kind of being an agent for the member, and it can do lot of things for the member, without member having to kind of ask. So this is where you can say you can set up certain actions. It can be an agent for the members and can do a lot of things. And then from the, financial institution perspective, you can be the bank GPT kind of a provider. So you can become similar to chat GPT, a bank GPT where you are holding the relationship with the member. And this bank GPT obviously is integrated with your stack, with your institution, but also it can help you become a more of a marketplace of financial information.

So you can bring other partners, and allow the customer to have a consolidated interaction with the financial life and also be the, guide in terms of, financial wellness and so on and so forth.

So this is all feasible combining generative AI and agent AI. And this is a journey we are at. And this is where, you can see like a dramatic dramatic difference.

And you have to rethink what online banking is. You need to rethink what the call center is. And, I’ll turn it over to Cortes, to show us a demo of where we can go with this agentic AI and generative AI approach.

So Cortes, if you don’t mind, again, can I remind you, Cortes is gonna show you a demo of SPHER Nexus Orbit, which is something we are working towards? And, I’ll turn it over to Cortes to share his screen and do the demo.

Awesome. Awesome. Alright. I’ll ask for Laura from our support team to kinda jump in here, and we’ll walk you through a demo of this next journey.

Right? The iteration, I think that Jack articulated was was really spot on, you know, going from conversational AI to potentially now leveraging the intelligence of generative AI. And then lastly, kind of at that top quadrant, potentially now adding in the agentic AI functionality, and what that’s gonna do in the digital experience world is absolutely amazing. So here’s a I like to call it a visual road map.

We’ll start with Orbit. Orbit is really meant to be a bolt on solution to your existing online and mobile banking. You can see at the bottom right, I’ve got a chat bubble that’s gonna, allow us to begin the conversation.

So let’s start with something simple. Right? Most chatbots today, we know, can do easy things like FAQs and routing numbers. So I’ll ask, you know, what’s my routing number?

It’s quickly gonna evaluate that information. Really easy, straightforward. We could dive into that. So let’s go a little bit further.

Leveraging now some of this agentic and generative AI technology.

I’m a student. I’m looking for a credit card.

It’s gonna come back with maybe some follow-up questions to help me dive into the specific features that’s gonna make this card and this offer potentially a little bit more tailored. So I’m looking to save money, maybe cash back, maybe I’m gonna be traveling. So let’s look at some rewards that’s tailored to my needs.

Awesome. Let’s go ahead and get started.

So now Generative AI is gonna dynamically search the website, pull in product guides, look at any of those third party data sources where AI now has access to to be able to come back, highlight specifically on the site exactly the card, exactly the offer that’s the best fit and best suited for me based on that knowledge, based on that background about me as the user. It’s highlighted exactly the features that’s gonna make the most sense based on the context it has about me and the relationship I have with the institution potentially, and then it’s gonna pull in potentially and highlight those top five reasons in the chat experience.

The great thing here is now making it a next step further. If I like the offer, I can hit yes. Let’s apply.

It’s gonna open up a window for me to now upload all the appropriate documentation to execute and apply for the loan right here, all driven through the conversation, all driven right here through chat. So I’m gonna upload my ID, maybe proof of residency, maybe proof of income.

And the cool thing here is to save time, to save energy, we’re gonna process those, leveraging now that Agentic AI functionality, and execute upon this application by extracting the information directly from the ID, directly from the material that’s been added. So let’s go ahead and hit apply.

So it’s gonna grab the information. It’s gonna pre fill the app. It’s gonna upload everything here, and we’re gonna connect to the API. This could be an API over to, MeridianLink, SWBC, Terafina, FICS, or Origins, just as an example. And all within ninety seconds, I’ve completed the loan application. I’ve processed it through. Really simple, really easy, straightforward.

You get back an automatic approval or potentially a application status update right here on the screen. Really easy, guys, to leverage this technology and this lending workflow. Obviously, very configurable, and we can put disclaimers and speed bumps where appropriate.

Alright. Now let’s log in. Let’s go one step further. So now we’re inside of digital banking. Imagine we’ve already authenticated with within the app. You can see I’ve got my accounts here. We’ll start with some basic functionality.

Maybe the first thing we’ll do is we’ll transfer some funds. So you’ll see how we can do transactional workflows now within digital banking.

So I’m gonna move some money from my savings to my checking account all right here within chat.

I can do so with voice, by talking to it, or typing it in just like we’re doing now. I’ll say yes. Let’s go ahead and do it.

You’ll see the balance will update. There it is. Fantastic. Again, leveraging chat, leveraging this conversation window to execute your banking.

Let’s go one step further.

Let’s maybe take a look at our streaming services.

Let’s look over the last year. Let’s see how much we’re spending. I got two teenagers in my house. I don’t know about you guys, but it seems like I’ve got, you know, one or two Hulu accounts. I’ve got a Disney plus account.

God knows what else. So let’s do an evaluation of all the accounts I have here within the institution and take a look at some analysis.

Let’s do it. So it’s quickly gonna fetch the data, look across my bank statements, look across all my recent transactions, all of the accounts I’ve got in here in the mix, and it’s gonna come back with a full analysis of how much I’m spending on streaming services over the past year.

Love it.

Oh my god.

Worse than we thought.

You know, six thousand dollars for a total expenditure on streaming services is just absurd.

So we’ve gotta get a handle on this. It’s really out of control.

So just, again, you know, really highlighting the level of functionality on how we can take agentic AI and generative AI and leverage it to really enhance your overall digital experience from a mobile and online banking perspective with Orbit, again, being that bolt on solution to your existing framework.

Alright. We’re ready to dive deeper. So, Laura, let’s go ahead and navigate over into our Sphere digital banking experience.

Sphere is now meant to be a stab at replacing your existing online banking provider.

With leveraging this ChattGPT like experience fully embedded into the digital banking environment, you have an advantage that’s like no other. Imagine having a financial assistant right there at your fingertips, Chad DBT that’s trained with all your insights about your banking relationship, again, being your in house financial advisor.

So let’s go ahead and open up that Sphere window, and let’s do something we know that digital banking just can’t do today.

I wanna plan and maybe start investing for my ten year old son. He’s going off to college soon, and, he’s a brainiac, smart kid, all honors classes. He’s a magnet and everything else, so I wanna begin to prepare ahead. Those tuition fees are getting really expensive.

So we’re gonna execute a plan and kinda help, you know, walk through maybe some savings options.

So our system is gonna provide you some questions and some context to help build out this savings plan.

Boston University is where I went for grad school, so let’s use that here for the synopsis.

I’m gonna click on my internal and external accounts that are at the credit union and maybe even at an outside bank as well. I’m gonna put a three thousand dollar monthly contribution, and let’s go ahead and start the investment strategy.

You know, once I get started, a couple things to note that are really important. You know, generative AI is so smart now where it’s gonna ask even some follow-up questions where, do you wanna add any other account? This is a big decision. It’s a lifetime move. So, yeah. Maybe let’s go ahead and add a joint account that I have with my spouse as well. It’s a great question here.

As we add the spouse account, you’ll notice one thing. This account’s at an external institution. It’s at Safe Credit Union, and typically within digital banking today, you’re gonna be asked to now have an interaction, an interface in place from digital banking over to a third party data aggregator like MX, Plaid, or Yodlee. But with Agentic AI and with Interface AI, we can log in one time to Safe Credit Union, and our technology is gonna capture the information.

We’ll go ahead and log in. It’s gonna capture the information from Safe Credit Union one time, and it’s now gonna pull in and extrapolate that data and include it now within our digital banking environment without the need of a third party data aggregator provider. One less third party. One less vendor to manage.

So great. I pulled in the spouse account. We’ll go ahead and hit continue, and let’s finish off the investment plan. You see the Safe Credit Union account now is visible. It’s now there. Awesome.

Let’s go ahead and execute this plan.

The cool thing here, you know, Jack mentioned this early on, and it’s really an important part, is, you know, the asynchronous design of everything that we do, you know, allows you to walk away. Just like Chatt GPT, when doing that deep reasoning, we can execute this plan, walk away, go do something else. You’ll get a notification later when it’s ready to go. But right here in real time, we’ve got the full synopsis of this complete savings plan, this five twenty nine plan to get junior ready to go to Boston u.

Looks like I’ve got a quite a quite a big, investment ahead of me. I may need to pull in some additional data sources and maybe start a side business to get ready for this. But we’ve included all those external and internal accounts in the evaluation.

We’ve even looked at third party factors. The AI has done all the research, taken care of all the legwork, but here now is where the real power comes in. We can execute this plan for the user, again, saving them time, saving them energy, saving them money, by just directly going out to the core, opening and funding the accounts, setting up those preset biweekly, weekly, or recurring transactions to move the money from the savings, checking, or even those external accounts to be able to fund and execute this plan. Really awesome.

The great thing here is we can also share this plan. This can turn into a PDF, can turn into a document if you’re not ready to execute. This is obviously just a highlight of the functionality that digital banking with the intelligence of generative and agentic AI can really do.

Again, one example, but, you know, just take a step back. Can you imagine this being leveraged in the lending scenario to really revamp what that lending process is, mortgage process could be? Can you imagine this even being leveraged in a car buying experience, amongst other things? Again, really exciting.

Just wanted to give you a quick sneak peek on what that digital banking experience would look like as it’s been revamped now with Agentic and generative AI. Alright. One more example. So, Laura, let’s go ahead and dive into NexSys.

NexSys is our frontline assistance tool. This is meant to completely transform the entire call center operation.

You know, we heard during the, you know, q and a session, the back and forth with Todd and and John throughout the call and the webinar today around, you know, how important it is to potentially have that one AI solution, that one AI brain from a voice and chat and conversational experience, and now potentially even on the employee AI side. It’s gonna continue to help us drive efficiency, but NexSys is gonna take things just one more step ahead.

So let’s do maybe two or three use cases here. We’re logged in as Samantha.

Samantha is a is a call center staff employee. She’s sitting at her desk. We’ve got API connections to all the third parties that she’s gonna need to interact with on a daily basis, you know, from Fiserv or FIS to Origins to Core, you name it.

And AI is gonna be the first line of defense taking care of all the steps that are necessary to help drive transactions and to continue to help drive automation away from the call center, away from CCaaS, and handle it all in house, again, with AI first.

So as you can see here with this first example, we’re looking at a chat where AI and the conversation from the person have been going back and forth. It’s been helping them drive an interaction around a wire transfer.

Looks like we’ve got a user that’s looking to send a wire, maybe around five thousand dollars or so.

AI has already completed six of the seven steps that’s necessary here to execute this transaction, and Samantha’s just being brought in as the human in the loop just to finish it off with that final step. So let’s go ahead and put in our password. Let’s log in, and let’s complete that last step in the process.

You can see it’s a fully authenticated user. It’s a known payee. It’s a repeat transaction.

Low risk. It’s come back with the score. Everything looks good. I’ve got three green check marks. I can quickly hit authorize, and now that is out the door.

Samantha had all the context of the previous conversation.

We were able to prick you know, quickly bring up, right here on the screen the core experience or maybe even the payment system just so she can validate things if needed. And just like that, we were able to get that wire out the door with that last step. Okay. Our next use case here, looks like there’s gonna be a callback request.

Maybe this is a mortgage. Maybe this is a small business situation where you want an expert to be ready to be on the phone, and maybe Samantha’s just that right person to handle this. This person could have been communicating with the chat throughout the day or maybe earlier, and they’ve scheduled this for later in the evening. Fantastic. Looks like they’re ready to go. We can call back the customer.

We can connect to Coconut or any other third party system that you’ve got out there that can, you know, coordinate those phone calls to those various specialists that you may have, rather in branch or maybe externally, let’s say at a mortgage center.

Great example.

Alright. We can go ahead and end the call, and let’s go to this last use case.

Samantha’s gonna be pulled in here again as that human in the loop, where AI has already handled the majority of the conversation. All the context, all the research, all the reasoning has already been done, and now Samantha can just pop in and just finish the final validation.

Looks like this is just a credit limit increase. This is a user that we already have today. They’ve got an excellent credit score, but maybe there’s some documentation missing. So let’s go ahead and request the income docs.

We’ll click that. It’ll pass that information request back to the chat. Again, AI and human are working hand in hand here, making Samantha really intelligent, and it looks like we just need to request the information. We’ll scan the emails. We’ll scan potentially this document source or whatever it may be to be able to find whatever that piece of missing documentation here is.

We now have a up to date income statement. The request looks good. All low risk. Let’s go ahead and approve it.

Samantha is now effective at her job, being able to navigate through multiple APIs. From a technology and from an infrastructure perspective. We’ve handled all the legwork. Right? Making it easy for you to take one single employee with one single application, and we automate all the mess on the back end with those third party systems. Again, making your call center staff more efficient, reducing those call and those handle times, and allowing them to leverage generative and agentic AI technology.

Again, this is a visual road map, a visual presentation of where our technology is going, how we’ve started off from the very beginning of this that conversational chat. We’ve grown it with generative AI. And now with Agintiq AI, things are just really taking off. Orbit, Sphere, and Nexus just gives you that sneak peek of where Interface AI is going. So thanks for the time today. Jack and team, I’ll turn it back over to you.

Thank you, Cortez.

Okay. So my favorite part, q and a. And while we are answering the questions, if you guys want to have a customized demo, feel free to scan this and request, a demo, and we’ll be more than happy to show you the demo of all our solutions.

Okay. So let me bring up the q and a. So the first question is for Todd.

And this is a classic question. We pretty much get it for all the, webinars we do.

How do members like AI voice?

Do they, see a I mean, do you see members resisting? They wanna talk to human even though AI can do the job? So, Todd, what is the answer to that question?

Sure. The answer is yes. We do see some resistors, of course. Right? It’s human nature with change.

We see kind of the classic three buckets people are familiar with with changes. We’ve got, you know, a third that were early adopters and all over it. Another third that were kind of fast followers. You know, we give them education. They’re willing to try and use it. And the last third is kind of your classic resistor bucket.

And it and what we have learned is that we just want to roll up our sleeves and understand, you know, where does the resistance come from? Is it, you know, because I just wanna talk to a person? Is it because I’ve got situations that this won’t help me solve? There’s all different reasons for that.

So we try to get at the heart of what that is. And we’re realistic. We can’t solve that all for everybody. And we made a decision probably after a few months, that we were live to go ahead and mention, and you’ll hear this if you happen to call our menu.

Later on the menu, it says, if you just wanna talk, to a representative, you can interrupt me at any time and say a representative.

So we we let them know that and that that even in and of itself comes with some risk because we’re like, okay. Is everybody just gonna say representative and just and just bail. Right? And they didn’t.

So, you know, that told us, okay. We’ve got a, you know, a healthy contingent of people based on the adoption that are fine just using the system and they want to use the system. And then we still have that outlet though. It’s like, hey.

If I just need to talk to human because I know my use case is unique or I just I’m just never gonna be a fan of bots no matter what you say to me, we give that option. So we we wanna continue to meet people where they are, educate them, bring them along if we can. But if we can’t, we’ve got people for you to talk to. We would always will.

Okay. So, Todd, a related question to, the previous question, how has member sentiment changed post implementation if at all?

Yeah. Member sentiment. So, yeah, we have basically and, you know, we do NPS and these types of things. And and one of the things we found was that we needed more timely information from the phone system to really kind of understand, you know, where IVA is at.

Traditionally, we did, like, you know, quarterly NPS scores, and we would say, you know, did you happen to use the IVA? What was your experience like? It was too late, you know, in time and people are trying to think back that I happen to use in the last, you know, month or two or what have you. So we worked with Interface to just implement a very simple solution.

Sometimes simple is best, where we say, hey, are you willing to take a survey? I’ll text you a link. And if they say sure, then the bot goes ahead and text them a link. It’s a simple survey.

They can fill it online, just a few questions, and we’ve been getting more information. That’s a pretty recent thing we’ve done in the last, like, ninety days. And coming in that channel, then we’ve got a mix now of, hey. You know what?

This is working well for me. Or, hey. I’ve got this issue. I tried to do this thing and it didn’t work.

Or, hey. I’m one of those people that just will never like AI and the whole thought of bots. I just like people. So we we see all that in the feedback.

But we, again, are you know, just roll up our sleeves, try to find any avenue we can to get that information, and then work those channels where we can drive change.

Okay. So now this question is about fraud.

Have you experienced any fraud where bad actors are getting past the authentication method to access member’s account?

We have not. You know, to our to our knowledge, we have not had, you know, members complain that, hey. Somebody came over the phone system, you know, and accessed my information without, you know, my knowledge specific to that.

You know, we do leverage some of the tools, of course, that that interface provides like the base package, and we are in the process of implementing the more sophisticated caller ID forensics, because you also don’t know what you don’t know. Right? So, no, we have not had any big, you know, fraud experiences since we’ve gone with the IVA.

Okay. Okay. So, Todd, this is a question about, kind of getting adoption. So what kind of instruction training tools do you use to help your members adopt using AI?

Yeah. Absolutely. So when we went live, you know, one of the things we did prior to go live is we said, hey. Our phone system’s going to be changing, and we’re gonna introduce you to this new capability.

And here’s some common things that this, idea can do for you. You know? So we we gave them basically a nice glossy short, you know, cheat sheet if you will of, hey, I can move money. I can check a balance.

I can do that stuff just with a few key phrases. So it, a lot of early education and that helped, you know, to some degree. We post things on our website saying, hey, if you wanna use the IVA, this is what can do. And we keep kind of a running list there of the things that it that it’s capable of.

So a fair amount of education. And then when people come through to the call center, we’re always encouraging our call center staff that if they if somebody comes through and says, hey, you know, I wanna move a hundred bucks from checking to savings, that they’re taking a minute to say, you know, next time if you’d like, you can do this yourself.

You know, and you can do it in the phone system and it’s super easy and here’s all you have to do. So, you know, we we use those different avenues, just to continually educate, you know, the members on what’s possible and and because it’s dynamic and we’re we’re always adding new intents and experiences. So we need to have constant education in that regard.

Okay. So the next question, I think you kinda answered it, but let’s make sure the answer is clear. So how does your monthly call volume for phone banking compare to your old phone banking?

Yeah. Yep. So, yeah, we’re we’re up from like, you know, fifty seven hundred, fifty four hundred a month, you know, prior to go live to eight thousand calls now, you know, eighteen months later. And and part of that, of course, is because we’re growing as an organization, you know, etcetera.

And it’s and so it’s been extremely beneficial that as we’ve grown in that kind of call volume, we have not added to staff at all, in our call center because we have, you know, the the level of automation that we do. So with us, because it was a small call center to begin with, it wasn’t about efficiency and replacing staff. It was about being able to scale. And we’re we’re absolutely seeing that come to fruition.

If we did not have this, we we, of course, would have had to hire additional, you know, call center staff to even come close, you know, to the ninety four percent, you know, answer rate, you know, and the and the one minute hold times that we have today.

And we’ve had success because we freed we freed even with the increased call volume, we freed up some capacity. And we’ve done things like, transition ITM support, so interactive tele machines. That support which used to be handled by branch staff is now handled by the call center because they had, capacity to do it, with the same five people, you know, six people that we had prior, you know, to this go live. So we have seen those benefits and we’ve been able to to leverage that additional capacity in in new ways.

A lot of people don’t realize that, there’s a huge amount of abandonment rate because the wait time is so high and so on. So they say, hey, AI has come in. We should see a decrease in volume, but initially, probably doesn’t go down because more people are able to get in. Right?

But the natural curve will bend because of the AI automation, and the customer member experience will improve dramatically as well because wait time goes down. They are able to actually reach you and so on. Right. Okay. So then the next question, I’ll answer it. This is about, we currently use Genesys for our IVR.

Would we be replacing that? So Genesys obviously does IVR, but also it is your full kind of ACD system. So we don’t replace Genesys completely, but you don’t need to use the IVR function of Genesys.

Our technology will basically be the IVR replacement. Like, Todd replace their existing IVR with AI voice, and they continue using RingCentral, it’ll be very similar with any, technology. So, we can sit behind IVR, but we highly we don’t recommend it. It doesn’t make sense.

And, generally, you want to replace your IVR with the voice AI kind of technology. Okay. Other question is kind of technical. Is the biometric authentication done through an integration with online banking? So typically, the biometric integration is, done with whatever your source of truth is for the phone number. So, generally, that’s the core. So we will send a link to do the biometric authentication to the number associated with that account.

Todd, do you wanna add anything here?

No. Yeah. I mean, that’s that’s spot on. So, yep, it’s gonna look at the primary phone numbers that, you know, that you have in the core, and that’s what it’s gonna send those authentications to. Yep.

Yeah. Yeah. Wherever your source of truth is is where you will do the integration for that.

Mhmm. In general, we we can do integration with, online banking, especially relevant for, AI chat. So in the AI chat model, we can basically do integration with the online banking Or you can authenticate them separately from online banking depending upon, your, online banking vendors support for the single sign on kind of an approach.

Okay. Then the next question is, is the learning model separate instance for each FI or combined with other financials?

So we are learning from all the impression we are having across the FIs.

But we are not using, kind of identifiable data from different FIs. It’s purely for learning. So we do learn from all FIs, but it’s not really kind of identified and, and stored in in that way that this is the learning coming from this FI. This is more like, okay, anonymous learning from different app, FIs.

Okay.

So this is again a technical question.

AI has come a long way since early chat GPT versions, but it’s still a concern out there that this AI solution can make errors. How does it interpret the address error rates and eliminate any errors that are made?

Yeah. So in general, the chat GBD kind of technology has improved a lot. You might have seen that in your kind of the, kind of the personal life, experiences as well.

Then from the interface dot AI perspective, we have invested a lot in kind of grounding the answer into the your knowledge.

There are various kind of technologies and approaches you can use to do that. And, our answers are, kind of better than human accuracy in terms of business and integration.

Can AI make a mistake? Yes.

But we are seeing that the right of error is actually less than a human making a mistake. And actually AI can act provide much more consistent answer than kind of the human in the contact center given that there was short tenure, the and there’s a lot of turnover and so on. So in a way, AI performs better than the human in the contact center.

And then also there’s an up and down kind of, you can reward the answer up and down. So we learn from that as well. And, of course, there’s a huge amount of QA you’re doing before you’re going live.

So, Todd, anything to add there?

No. Like you said, there you know, you just do a lot of QA to submit you know, test all your use cases, your edge cases, all those types of things. You’re you’re, of course, gonna find things, and that’s time well spent.

Okay.

Believe it or not, I am out of questions and we are out of time at the same time.

So really appreciate, our panelists. I thank you, Todd, for taking the time to join us. You, Todd.

Thanks for having me.

And, again, thank you all for attending on a summer Friday. Really appreciate your time.

Reach out to us if you would like to continue the conversation, and, we’d love to engage with you guys. Thank you.