← CIO Leadership Live
CIO Leadership Live
From Paper Checks to AI: How Deluxe Is Reinventing Itself as a Digital Payments Company
00:24:52
Loading summary
Transcript35 lines
- [00:05]
A
Welcome to CIO Leadership Live. My name is Lucas Marion. I'm a senior reporter for Computer World magazine and I'm here at the CIO100 symposium and awards show. And with me today is Yogurt Jayapakasam. You are the Chief Technology and Digital Officer at Deluxe. It's a business services firm processing payments marketing, promotional products, check printing and financial technology solutions.
- [00:29]
B
Yes, we are.
- [00:29]
A
Thank you so much for taking the time to talk today. I appreciate it.
- [00:32]
B
Absolutely. Thank you so much for the opportunity.
- [00:34]
A
Absolutely. Now I'm old enough, I remember Deluxe as just check printing. So when you say that name, that's what I think of. But you, you've really diversified in the industry because you've evolved so much over the past decade from legacy check printer to tech enabled services firm. How do you see your role in in driving transformation?
- [00:59]
B
That's a fantastic question. In 2018, our current CEO took their role and the job of his role was try to clarify the long term strategy for the company. And what we realized in the process is cheque is the first alternative payment mechanism after currency. But the companies for so long focused on manufacturing as the business identity, not necessarily making alternative payment as our business. So that's the strategy clarification that our new CEO brought in. So in 2022 when Hai was approached for this job, the primary need of the job is as a manufacturing company. We use the technology as a supporting mechanism to drive the business. But now we want to pivot to becoming a payments and data company where technology becomes the product and services that we sell and not the physical paper checks that we used to produce. So my job in the company is actually driving both technology and cultural change in seeing technology as the product and service that we build and deliver, not just as using it as a support mechanism as well. So that's basically what we do inside the company. And in the last three years we drove significant transformation inside technology using technology but also across the company in driving clarity on out of the $2 billion revenue we make, one still comes from the print side of the business, but the other 1 billion comes from merchant services, B2B payments and data driven marketing solutions all underwent a massive technology transformation.
- [02:51]
A
How has AI played a role in that?
- [02:54]
B
So I see AI's role as a force multiplayer. I have been in technology business for over 25 years now and have witnessed many technological transformations starting with agile transformation, to cloud transformation, to big data transformation, to platform driven transformation. I see AI's revolution now is culmination of capitalizing all the technology transformation that already took place and now they are all converging with AI. And I see this as an opportunity to. Initially our focus is using AI to drive cost optimization, customer support, internal employee experience improvements. But the real work of AI is going to be fundamentally reimagining everything we do inside business. And that's what is exciting right now for me.
- [03:51]
A
How do you build or how are you building a digital first culture inside a company that historically has operated in more traditional ways?
- [04:01]
B
That's a really meaningful question for me because until I took this job, mostly I saw my role as a technology role and I focused heavily on what is trending in technology, what is the most cutting edge architecture that we should be creating, how do you create more cloud, native solution? It's all important, but they are not the one that is going to help you succeed. So end of the day for company to truly succeed with the transformation. Technology is the smart smaller part in my opinion. It is mostly people and culture. So what I mean by that is understanding where your company comes from, what cultural foundation that you already have, acknowledging that and leveraging that to your strength and then driving the transformation forward is what we try to do inside the company. So let me try to explain what I mean by that. As a manufacturing company, our company was and is very deeply rooted into customer commitment, service, quality and trust as the foundational elements. So driving a most complex project that I would find it very difficult outside, I found it relatively easy inside our company because our people are so committed to no matter how hard the project is, doing the right thing by our customers. So leveraging that foundation, shifting it to digital transformation, it became more of teaching them why the transformation we still drive into the digital world is still going to protect those values of customer centricity, driving trust and caring about our community. So eventually my role became more of combining technology with HR and transformation together, which is basically how we continue to drive it forward.
- [05:58]
A
Check printing, still financial services function. How do you handle the sensitive financial and marketing data? And how are you leveraging AI and analytics to create real time value for your customers?
- [06:12]
B
That's again another question. Every company has to realize that your data is your real asset. Technology comes and goes. Truly what you have is the data that you have been building with your customer. So the approach that we have taken as part of our transformation is what I would like to call platform driven approach. So we chose to build three different platforms for the company. That's Delectra Data, which basically is to bring and organize companies entire data that used to be Siloed because of multiple acquisitions that we have done into a well structured domain driven data model. So in simple business term, what it means is we no longer look at data as system specific thing, but we see that as a company specific asset. So all the customers that might be doing business with one part of our business now becomes part of one data domain. So we now know 360 degree view of the customer, what product they buy and why are they buying it. And then you can also try to understand the various type of products that we have, what stickiness it is creating all of those things as well. So the first foundation is creating the domain driven model in organizing the data. And at the air we use AI heavily as well. So in my previous job when we had to create a data lake, it took us more than two years to create the first version. More than half of the time was spent by subject matter experts who understood specific business unit data elements. So they were heavily involved in trying to map those data elements to the new data structure. But with the invention of generative AI and large language model, we were able to use large language models to help us to do the same mapping. That rather would have required subject matter experts, domain expertise and months of their time to drive this forward. So that was the first breakthrough. But after that it is basically applying governance on what is permissible, whether you are using this data for marketing. But even there, is it, is it credit related data or is it generally promotional data? So applying more metadata driven permissible data rules and on top of that applying additional AI driven insights as well. So inside the company we look at applying those AI capabilities through three different lenses. So we apply AI for technology which is using AI to change the way we deliver software product. Second is AI for business operations which is using AI and our internal data to modernize and automate what we potentially do with every department. For example, finance doing month end close used to take weeks. So month end close supposed to be month end. So it would rather take a couple of weeks for us to close the data. Now with modern data lake and AI, we are able to close it within a few days and no longer take weeks. So the same way operations team is able to automate exceptions and various other things. So all pieces put together, it's organizing the data, applying the right governance model and rules and then really looking at what business use cases that you solve through the lens of AI for tech, AI for operations and AI for customers.
- [09:41]
A
And AI helped assist in doing that?
- [09:43]
B
Absolutely.
- [09:44]
A
That's remarkable. So you literally. Because typically it's Garbage in, garbage out. So organizations need to clean the data first before they want to apply the AI to it. But in this case, you use the
- [09:55]
B
AI to clean the data as well.
- [09:58]
A
Hadn't heard that one before. Interesting. Do you see AI as playing a larger role in Deluxe's offerings, perhaps in fraud detection, payment risk, customer experience, personalization?
- [10:12]
B
Absolutely. In the pillar that I talked about, A, for customers, it is all about leveraging AI with the way we deliver our product and services. So the example that I can give you with the service side of things, as part of the platform we built, we call it Deluxe AI.
- [10:31]
A
What does it call themselves?
- [10:32]
B
Deluxe AI Deluxe. Okay. So it's an internal AI platform that creates wrapper around the public LLMs and then applying our own retrieval augmented generative business rules on top of that. And then we created a generative AI chatbot for the company, but then that gets customized for every business unit. So we have one chatbot, but it has the intelligence and awareness of every different business unit. So it appears like each business unit is having its own chatbot. So that's how we started to handle the customer service. So our merchant services has their own version of the DAX, what we call Deluxe AI Assistant, we call it DAX, the same thing for merchant B2B payments, we have another version that gets out to the customers as well. So that's basically more and more of how we are streamlining the customer service part of it with generative AI, but from the end customer standpoint, we have receivables automation platform, it's called R360 Plus. But the way we are using generative AI, there is to apply cache application detection module and even extracting invoice data and customer payment instrument information to streamline the whole process that rather human analysts would take hours and weeks. Now we are able to automate and bring more actionable intelligence into that capabilities.
- [12:03]
A
Is the AI that you deployed a cloud service offering or is this internally developed?
- [12:08]
B
So we use a combination of both. So the public cloud services that we use, what we get access to, which is typically all the publicly available large language models you get access to, which you want to take advantage of immediately. But then we built our own what we call Deluxe Trust layer that basically prevents what data that we need to prevent within Deluxe reaching to the large language model, which is where the RAG or retrieval augmented generative architecture comes into picture. And then we also train some of the large language models with our own data, which then gets a flavor of our own personalization, if you may, has
- [12:51]
A
RAG let you down in any ways.
- [12:55]
B
See, the common challenge with RAG is hallucination. And we built enough testing frameworks to make sure that our architecture is well protected enough. So far we haven't faced any challenge. And also the architectural approach that we took is more risk based approach. So we start with no risk type use cases, which is applying RAG for internal, AI for tech and then get that trained and then you go up all the way to the customer facing applications, which is eventually how we delivered what we called Deluxe AI assist.
- [13:35]
A
Okay, what sort of safeguards or guardrails have you put in place in terms of data security, accuracy, ethical use, especially when applying AI to your business services?
- [13:49]
B
I think that is as much as data is your asset and fuel, it is also your most riskiest thing to protect. Because as I said, trust has been one of the foundational element of Deluxe and us being 110/year company, various financial institutions and banks trusted us with their data and their capabilities as well. So the approach that we take is we treat customer data as much as more critically than our own data. So we are not only looking at the permissible use of the data, we are also looking at what is right by the customer when we use this data as well. So it comes in many forms. So it begins with what internally we have called compliance by design approach, which looks into every use case that we have to apply to see are we applying the data within all the lenses of the contractual obligation to legal and privacy and everything else. But then at the senior most level we also defined AI executive committee and it includes our cfo, HR legal head, myself from technology head of operations. And the goal of that committee is we are not only leaving AI permissible use to the team, but on a frequent basis we are discussing how we are using and the various type of use cases team is solving from that executive level as well. But eventually, if it is a customer facing applications, we have a dedicated client advisory board and on a regular basis we go in front of the advisory board that has representation from many of our top customers. So we discuss how we are approaching AI and usage of their data and get their consensus before we actually use it and roll out products as well. So by applying multiple safeguards, I believe we are making sure that we are doing what is right by the customers that we serve, more so than just representing our own company's interest.
- [16:03]
A
Excellent. What are the biggest opportunities you see for Deluxe in the next three to five years? From a digital and technology standpoint, I
- [16:12]
B
think the biggest opportunity that every company has in my opinion is not only seeing AI and any other technological innovation as a cost optimization mechanism, but truly seeing that as a leapfrog opportunity. And that in my opinion we have a long way to go and in three to five years we have to upskill our talent. I seriously believe recently PwC published this research where they showed with the invent of AI the organizational structures potentially could become diamond shaped instead of pyramid shaped, the triangle shape. What it basically means is now if you think about diamond shaped organizational structure, that means entry level employees are going to be so minimum and the middle layer is going to be so wide and then the top layer will be smaller again. But the risk of getting into that there are already job reports that entry level jobs are shrinking, but the risk of handling it that way is that you are not going to have an entry level to train them to become your mid level career. And the current mid level career folks are already facing pressure of upskilling themselves while they are also dealing with their own full time jobs and family responsibilities and everything else. So for us to succeed in the future, in my opinion, many companies job is not going to be technology and adopting technology because AI is very hard to invent. But some companies are doing it really well, relatively easy to consume, but the real usage of AI is going to be are we able to retrain our people to reimagine how you would approach the problems that we solved in the last 50 years? Totally different way. And that if I go back to my comment about entry level career and mid level career, in my opinion entry level career folks are going to require more soft skills and business domain expertise as quickly as possible. But the mid career folks are going to require upskilling with the AI technology because they already have the soft skills and domain expertise and everything else. So how do you marry these two pieces together? And that as you can imagine is a huge cultural change management challenge. So it is as much of human resources challenge than technological challenge. And inside Deluxe we are partnering very closely between HR and technology to figure out how do we create joint upskilling program. It's not only focusing two teaching technology, it is basically focused on things that I talked about. Domain expertise, soft skill and bringing entry level and mid career people together so the hierarchy does not become a barrier and we can help them to learn from each other and help the company to advance forward.
- [19:13]
A
Just briefly talk a little bit about some of the soft skills that you're looking into now that you require for this.
- [19:20]
B
So I did a lot of research and I recently published the top 10 skills. But I'm not going to list all top 10 skills, but the top three would be the number one is ability to ask questions. So for example, if you look back in every company, customer acquisition, customer cross sell, targeting, customer service, all of that are the same problems that we are all trying to solve for many, many years. But every time technology changes, then you go back with, oh, we came up with mobile channel, web channel, even telephone and text messaging. And now in the age of AI, how would you approach customer service? So that is what I mean by asking the question or even reframing the question that you already have. So that I think will be number one. Soft skill. Yeah. Number two would be critical thinking. So not just blindly building on something that you already have and then making some minor change. But then how do you really challenge, is that the right thing to do? Is that the right way to handle that as well? Then the third one in my opinion would be deeply falling in love with the problem and really trying to figure out what is the outcome that you are going after. Then you, in between everything else, you should be willing to completely challenge it. So it is asking the right question, introducing critical thinking and falling in love with the problem deeply without attaching yourself to the solution that we created.
- [20:55]
A
Yeah, it's interesting because I've written a lot about soft skills, hard skills certifications, that sort of thing. And I thought it was interesting that you talked about the different levels of career and where the soft skills and the hard skills are going to apply. So the reskilling of middle management, the soft skill development, or at least looking for those candidates for the entry level employees, what do you do with the executives?
- [21:22]
B
That's a fantastic question. So actually inside our company we are developing what we call three by three. So it is AI skill level from beginner to intermediate to advanced, far entry level to mid career to executive. So we are not leaving executives alone in the training program. And I personally started with my peer group and the CEO on how we all should be thinking about AI. So when it comes to executives, my current research and the hypothesis is a smart executive who is usually connected with good network of executives, you always know which technology to take the bet on. And I'm pretty sure every CXO is talking about AI and how they want to invest in AI and what is the outcome they are trying to drive with AI. But the real challenge for executives is how do you really understand the technology and then use the technology to even reframe. How would you be approaching the business problems and business challenges inside the company. So for example, before AI, if you needed to have a strategic point of view on Genius act impact or stablecoin impact, you would be hiring some three letter big consulting firm, pay at least half a million dollar, give them three months for them to come up with a paper. So now you can actually go to Perplexity or even go to ChatGPT. And it's deep research mode can bring you reliable, really meaningful first point of view on what is the impact of Genesect. Not just generic, but it can also give you a point of view on what is specific for your company. So I actually took this as a personal challenge and I developed a research paper on Genius act for our company. It took me less than two hours, but I shared it with my entire executive team and the CEO. None of us were able to poke holes with the report that AI produced. So the point that I'm trying to make is we can't leave it to say that oh, executives don't have anything to do here. Now you can actually have a strategic point of view in minutes. What are we going to do with this? So how do we reframe that? So that's number one. Number two, most important in my opinion is see we all live in capital society and we will continue to have pressure for margins and pressure for growth. But how are we going to balance between driving business growth and bringing our people along in the journey? And that in my opinion is the biggest challenge. So in the three by three, what we are trying to do is how do we help each other in the executive team on the power of AI, but also the accountability of the leadership in using that AI along with humans.
- [24:27]
A
I can't thank you enough for taking the time to talk about these things. You had some great insights and I've really enjoyed this.
- [24:34]
B
Thanks so much for the opportunity. I really enjoyed it as well and
- [24:37]
A
I hope you enjoy the rest of the conference.