Transforming Banking Data into Customer Value

Jim Marous (00:16):

Several key data trends are transforming the competitive landscape of banking. First, the emergence of agentic AI is prompting financial institutions to reconsider how they organize and use customer data. Second, vendor consolidation is compelling banks to take greater control of their data strategies. And third, the growing sophistication of AI models requires higher quality data to provide personalized experiences while addressing risk and other types of biases.

Jim Marous (00:51):

In this episode of Banking Transformed, Don Permezel, Vice President of Data & AI at nCino joins us to explore how banks must evolve beyond mere data collection agents to unlock more customer value through AI powered personalization.

Jim Marous (01:10):

This discussion reveals that while many banks excel at data collection stories, they often fail to turn this data into actionable intelligence that enhances customer and organizational value.

Jim Marous (01:24):

Data has long been considered the new oil of the digital economy, but this valuable resource remains largely untapped for many financial institutions. What challenges stand in the way of transforming data into customer value? And where should financial institutions begin this journey?

Jim Marous (01:44):

So, Don, I'm really happy to have you on the show today. Before we start, can you share a little bit about your background and when you've been at nCino?

Don Permezel (01:51):

Yeah, of course. Delighted to Jim. So, if you can't tell by the accent, I'm Australian by extraction from Melbourne, Australia, where I worked for nCino for several years heading up their product development and engineering effort out of Melbourne, Australia, where we got a good sort of 50 or so staff in that region. But as of about six months ago, I was asked to immigrate to the United States to head office at this stage in North Carolina to lead up the Data and AI project.

Don Permezel (02:21):

So, that encompasses data platform, data engineering, data science, generative AI team as well as business lines in essentially pulling stuff out of documents with artificial intelligence and editing, based out of Melbourne, Australia, funnily enough – I took that along with me.

Don Permezel (02:39):

And the portfolio analytics team, which is really helping banks and credit unions understand their risk and return profile. And then very interesting business in the incentive and analytics team that helps operationalize and automate incentivization for bankers as well as visibility into mortgage application flow. So yeah, delighted to be in the position. Love nCino, and delighted to talk to you, Jim.

Jim Marous (03:03):

Well, it's interesting, when I look at my history, it goes back 45 years in banking, and we were talking data back then. We were looking at seeing how can you use data to build better customer experiences, but it's been a whole lot more talk than actual action and deployment.

Jim Marous (03:20):

And we all know that financial institutions have such a vast amount of data available to them, both internally and externally. But it goes without saying that many banks struggle to build both a data-driven culture and using data to do more than just know about the customer, but to deploy solutions that the customer can actually know, you know them.

Jim Marous (03:47):

What do you see as the primary organizational barriers that really stand in the way of going from reports and line items to actual deployment and value?

Don Permezel (04:01):

So, it's a great question, Jim, and I think you're right, that sort of the concept of data and better leveraging of data isn't anything new, and indeed, existed far before the AI revolution or large language models come around, whatever way you want to think about it.

Don Permezel (04:18):

But I see as well as the headwinds, which I can speak about a little bit – I also see indeed the opportunity for a little bit of a new era. Because what I've seen as I've seen the technology evolve and the opportunities to use that technology at different parts of the banking process.

Don Permezel (04:36):

I've also seen a cultural shift as well within banking where while all the problems are not yet solved, which I guess is what your question is about, and I can speak to those briefly, I see a renewed willingness and vigor to confront them, a determination to sort of unlock the power of not just large language models, but intelligence more broadly fueled by the data.

Don Permezel (05:01):

So, something else in your question that really resonated with me was, people sometimes mistake opportunity and data for simply the quantity of data itself and sort of having the data per se. But to your point, it's sort of almost like if you've got an oil field, it can't actually make your car run unless you get it out of the ground, and you got to get it out of the ground in the right way to create actual value from it.

Don Permezel (05:32):

And so, that's I guess, where the difficulty lies. So, there are a few elements to that. So, firstly, essentially, the data platforms and the technology and data lake, which these financial institutions use to access the data, it's not quite as simple as just plugging in a Databricks or a Snowflake or a what have you.

Don Permezel (05:54):

Actually, like a lot of effort needs to go into, well, firstly, I mean, it sort of goes without saying, but the information security environment and controls and all these sorts of challenges that can be pretty meaty engineering tasks. But also, the question of the cleanliness of the data and being able to have it knit together in such a fashion that it's able to be used by data science and product people to create real value for customers.

Don Permezel (06:24):

And look, that's a difficult challenge. I can speak about some more if you like, but one where nCino is not through any great genius or anything, but happens to be relatively well-positioned by providing this sort of single platform for commercial and other lines of business.

Don Permezel (06:44):

The other headwind that I would say is essentially the juncture of data expertise and banking expertise. And where I see banks perform really well in the space that you're asking about, it's when they're able not to isolate the data questions and the data analysis to a sort of ring fence that works with the potentially not so clean data and sort of tries to generate insights.

Don Permezel (07:14):

But when the culture of the business shifts such that that data obsession and data-driven decision making is woven much more deeply into the organization, so that analysts and the data scientists are actually asking the questions of the data that you need to go and create that customer value.

Don Permezel (07:34):

So, it really needs to be integrated. You need pretty heavy data platforms with very well-thought through data pipelines, being able to supply that. You need a strong data governance framework, and nCino probably knows that better than anyone (as well as anyone else, I should say) given our client base and how seriously they treat the issue. And then you really need that juncture of analysis, data science, and business expertise to make it real, I'd say is some of those headwinds.

Jim Marous (08:07):

You referenced the work that nCino does to help financial institutions in this way. And the financial ecosystem overall has changed so much to the point where organizations now make a lot of decisions between the build versus buy.

Jim Marous (08:23):

It used to be that organizations kept all this intelligence, all this capability internal to their organization, but now, organizations of any size can actually build a really good data organization through partnering with third party organizations.

Jim Marous (08:44):

What internal data capabilities should financial institutions prioritize right now, either themselves or with partners?

Don Permezel (08:54):

Yeah, very good question, and a question of special importance in my role, and I think many roles across the industry. It's such a quickly evolving space, and you have to be absolutely concentrated on deploying your investment to develop the sorts of IP that make sense, and develop internal to the organization where that creates the most shareholder value and partner where the alternate is.

Don Permezel (09:23):

It might seem to be a relatively straightforward answer to your question, but the primary dimension that I think about when deciding whether to build, buy, or partner is really on the level of commodification of the underlying technology. And that level of commodification is actually growing.

Don Permezel (09:45):

So, it's not just that the underlying technologies are advancing, it's that the application of more general infrastructure is growing as well.

Don Permezel (09:53):

And an obvious example might be sort of modern and more recently emerging large language models, which are sort of commodified in a sense. As in you've got your big providers in OpenAI and Anthropic and others.

Don Permezel (10:10):

And the applicability of those models, they're performing so well that specific business problems that they previously wouldn't have been able to be addressed. They're then now because of the improvement in technology, they're able to be.

Don Permezel (10:24):

So, I guess that the answer is then as banks seek to … if they want to build, like fine tune their own large language model or build a random forest or genetic algorithm or whatever they want to do – should really be focused on their own bread and butter, what makes that bank unique?

Don Permezel (10:41):

So, maybe they're operating in a particular industry, maybe they've got a client set of a particular character, maybe they've got bankers, working under a particular organizational structure, building capability to solve the problems that are unique to them as a financial institution will always be a better deployment of IP than trying to beat Amazon or Google or Microsoft, or who knows, with a sort of a solution that's more generic.

Don Permezel (11:12):

Having this said, banks need to be very, very careful not to outsource too much such that they lose that internal expertise around data and AI. You can’t really outsource making the decision, owning the problem set, and owning the solution and the outcomes of the solution, because that just naturally sits with the bank, that can't sit anywhere else.

Don Permezel (11:37):

So, it's a little bit of this trade off where you really want to access all these global providers, especially as I said, in sort of genericized capability. But at the same time, you can't give away too much of the bank's own IP knowledge and expertise in the sense, because it needs that to flourish later on.

Jim Marous (11:57):

So, when you look at that, you talk about outsourcing the … I mean, the financial institution can't let go of the ownership and the compliance elements. So, the use of data, the issues of privacy and security, the use of the AI right now, there's all combining to build both capabilities, but also concerns within the financial services industry. And that's one of the things that we're seeing makes it so financial institutions sometimes hold back.

Jim Marous (12:28):

I mean, they fight with themselves because what they can provide to a customer from a standpoint of personalization and recommendations – many organizations take that old line thinking of no mistakes. They don't want to miss a mark on something, as opposed to realizing it's a balancing act of like a dimmer switch looking at what can I share with a customer? How can I make recommendations? Knowing that some of them may not be completely right.

Jim Marous (12:58):

How do you see (and this question probably would've been a different answer three and a half months ago or two months ago) the regulatory system changing or evolving as it looks at AI and data and privacy within the banking community?

Don Permezel (13:15):

Well, certainly, that's a good question. And just to sort of revert on the middle part before I get to the latter one, I think that with that cultural change we've seen in the industry over the past year and a half, two years, I think the point you made will be more and more deeply accepted throughout financial institutions globally.

Don Permezel (13:39):

That is the mindset shift as you very well put it of nothing can ever be wrong to … yeah, I like your analogy of the dimmer switch. Like how do we soberly look at risk? How do we soberly look at customer income impact, and how do we soberly trade that off? We'll trade it off or combine that with a view on shareholder return.

Don Permezel (14:02):

So, I think that maturity is something that banks are going through at the moment, and as they increase their, I guess, internal data and AI expertise, their data science teams analysts, and even sort of that expertise sort of filtering up even into the executive, I think you'll find a greater willingness to adopt more of the AI tooling.

Don Permezel (14:29):

But you're certainly right about that tension as well. I think you know people often talk about first mover advantage, but I think banks also feel a little bit about first mover risk, and sort of waiting for perhaps the others to jump first or getting the timing right. So, the first one jumps in the swimming pool, and you want to be the second one or something like that.

Don Permezel (14:48):

But this isn't anything new. Just to put right, I don't know if it's a plug for nCino, but nCino can sort of rightly lay claim to the fact that we were one of the big elements in taking banking to the cloud, and certainly yes, as you're talking about, moving into this AI world, especially vis-a-vis customer interaction, use of non-deterministic systems, et cetera.

Don Permezel (15:13):

Yeah, that's a big jump, and that takes a lot of guts. But you know what also takes a lot of guts? Having all your data in a server in the server room, and having that move up into the cloud. So, I think banks have taken these sorts of leaps before, and I think they'll do it again eloquently with artificial intelligence.

Don Permezel (15:31):

As far as the regulatory environment's concerned, I think you'll see as banks look to leverage AI to do things like take the waste out of their process to automate parts of banking that were never automated before, I think to unlock the banker to create better experiences for their customers.

Don Permezel (15:48):

I think the regulation will ultimately reflect from that and start to modernize around being more specific around sort of definition of models depending on which part of the banking flow that they operate in. I think things like the credit decision should be absolutely totally rigorous and will remain having the degree of governance, or perhaps even more governance than we see presently.

Don Permezel (16:20):

But other parts of the banking value chain as you said, I think regulation may modernize to encompass more of that dimmer switch mentality that you talk about providing that things are explainable, they're measurable, and it's demonstratable why the bank has adopted the policies and procedure, which it has. Which is interesting, I mean, I could talk about agent AI in that sort of framing, but I'll leave it at that at the moment.

Jim Marous (16:51):

So, Don, nCino has hundreds of financial institution clients. And each client, when they first engage with nCino probably says, “What we want to do is get our data in order. We want it to be clean data, we want to be able to use it, we want to be able to deploy solutions against it.” But there's obviously differences between your clients that leverage what they can do the most, and those that basically just have digitized what they were doing before they had this capability.

Jim Marous (17:26):

What differentiates those that are maximizing the value from those who aren't? Because the title of this podcast is transforming banking data into customer value, but even though you have the capabilities, not everybody does it. What differentiates those really strong success stories from the rest?

Don Permezel (17:50):

Yeah. So, a couple of small points I can pick up on that. So, one of which is your observation we see very strongly from nCino, and even sort of pre-AI revolution – those banks when adopting a platform like nCino to perform their banking, who instead of just as you said, sort of take what they were doing previously and then digitize it.

Don Permezel (18:16):

They've had less success as a rule than those banks that have said, "Hey, look, I'm going into a digital world …” and as related to your question, into an AI world. "I'm not just going to look at this in terms of a technology replacing my current process. I'm going to say, what should the process be, given the new technology?"

Don Permezel (18:37):

And I think that there's nowhere that that's more true than in the world of AI. Whereas, these capabilities become unlocked, it's those customers that only think about time-saving on a particular step, and those that think a little bit more broadly about the broader origination or rather major process and say, "Hey, how should I organize this differently to be more reflective of the technology?" It's the latter. It’s the latter that in general, do much better.

Don Permezel (19:08):

So, I guess I could build on that with one more step, which is where I see banks have a lot of success is when they are absolutely forthright about developing the tooling and analysis on their process. In some ways, it's as old as some of those, I think Adam Smith experiments with the I think it was coal miners or something like this.

Don Permezel (19:36):

But the modern version, the 2024 version of the time and motion study such that banks develop a rich understanding of where the variability exists in the process, where the cost exists in the process, where the value is created in the process, how the customer experience is affected by various process steps and importantly, where things go wrong in the process.

Don Permezel (19:58):

Something that whenever I talk to banking executives, one of the things that resonates the most strongly is just things going wrong. And how do you identify and be proactive and be one step ahead of things going wrong in the process?

Don Permezel (20:15):

So, these banks that really adopt this more modern and more scientific approach to understanding the science of banking, I call it, the process that creates the financial product – these are the ones that are also best equipped to sort of, I guess, respond to your question, in that they're the ones that are closer to their processes and closer to the measurability and the value that their process creates and where there is waste as well, that are also best equipped then to rethink their process in light of some of the new technologies.

Don Permezel (20:49):

I can give you one more tiny little bit of an answer just because I can't resist. I think what we'll see, and I think we're already beginning to see is with the new technologies, processes adapt more such that information collection is really pushed further to the left of the overall process flow such that the artificial intelligence machine learning statistical models (take your pick) essentially have more information through the process to draw more inferences.

Don Permezel (21:24):

And I think you'll just see that continue more and more, whether via deeper integrations and greater use of, sort of third party data providers. Or what I think you'll (well, I don't just think I know) see emerging sort of over the coming 12 months or so, you'll see just much, much, much more efficiency and automation around the collection of unstructured data.

Don Permezel (21:51):

And I think that's going to be so powerful that it's not just going to be, “Hey, we should rethink our process, so we give the AI and the machine learning more data points early on to automate through the steps.”

Don Permezel (22:06):

It'll also be, “Hey, since this data collection is now much more automated than ever before, why don't I collect more of it? Why don't I bring more data points throughout my operational process and even into credit decision world?”

Don Permezel (22:20):

So, I think it'll be an exciting time, and as I said, with the culture shift, I see leading banks absolutely taking the plunge.

Jim Marous (22:27):

You talk about that culture shift, and we see it in almost everything we're talking about recently in the podcast, is it's not the technology, it's not the back office – it's the leadership of the organization. Do they know what their destination has to be and are they moving in that direction?

Jim Marous (22:44):

And I've been fortunate enough to meet and speak with your CEO and it plays well in nCino as well, where that culture comes from the leadership as opposed to something that pushes up. It's really pushed down in the way that they view the industry, the way they look at the future and what's going to happen.

Jim Marous (23:02):

And when you look at the future, obviously, AI and agentic AI and all this personalization we're talking about the potential is expanding tremendously, almost on a daily basis. We can look at the newspapers and the news media over the last couple days and realize that basically it changes almost in real time.

Jim Marous (23:24):

What do organizations have to do from a data perspective to really be ready for what we're talking about here, which is the potential to build AI agents that will work on behalf of the consumer in the financial institution, or maybe outside the financial institution?

Don Permezel (23:43):

So, three obvious ones is, number one, what I spoke about a little bit before in a previous conversation, being absolutely serious about the data platform that the organization's using, the cleanliness of that data, the interpretability of that data, and the understanding of the potential of how to use it.

Don Permezel (24:07):

I think what's also really important is essentially the level of culture in the organization and a culture to be persuaded by innovation. So, something which we've done at nCino, and I think many other institutions are doing as they as they carry forward – as you say, the technology is emerging very rapidly, so you need to have a think about how you best leverage it – is that as business as usual is being conducted.

Don Permezel (24:37):

So, in nCino's case, as nCino is building software it might be in commercial banking, it might be in retail banking, it might be in portfolio analytics, a particular customer or set of customers desire a particular feature.

Don Permezel (24:51):

So, rather than go ahead and prioritize and backlog and cart up and build that feature alone, which is what we do as a software development company – right from the start of the process, the very first question is, how can I supercharge this with AI? How can I supercharge this with machine learning? What parts does the model fit in? Do I need a predictive model? Is there a particular statistical inference that I can make from this data to that data there?

Don Permezel (25:20):

And then through that sort of almost process change, as in that's a step in product development, is saying, how can I supercharge that – it's then the invitation then to work with the data science team, data science experts, data minded product people, et cetera, to be able to refine that so that it swims much more deeply in the organization.

Don Permezel (25:44):

The third thing I think is just the level of expertise. And it's only my experience, so throw it away if it doesn't make sense. But I've found that when it comes to data science and finding new applications for data in customer value, it's not necessarily about having enormous amounts of data scientists. It's about having a few very, very good ones.

Don Permezel (26:11):

And I think that really sort of finding the absolute richest talent in the space is key critical, but it's not enough. You need that pool of talent from a data science perspective, but then you need the broader business to have a level of fluency that they're asking the right questions around what products can be delivered with that data. And it's actually those two sides coming together where I see the most opportunity.

Don Permezel (26:41):

And I think over time, you'll find over the next 2, 3, 4, 5 years, you'll find a level of data expertise, even machine learning expertise, filter up through the organization into the executive.

Don Permezel (26:57):

So much so that I can imagine a future where you put a graph in front of the CEO and he or she says, “That looks great, but show me the statistic, show me that that is actually demonstrated.” And I think you'll just find that maturity happen over time with the greater focus on data as the culture is changing.

Jim Marous (27:18):

Yeah, we do seem to be catching up with ourselves often. It's a very interesting time in the banking world when you look at data and AI and the capabilities of what's there. And again, when you're looking at a regulated industry, it becomes interesting because you're continually looking at this power of what you have and the risks that may be associated with that. And I'm hoping as a person in the banking industry that we can balance that pretty well, as I said, the dimmer switch.

Jim Marous (27:47):

So, let's just take a short break here and recognize the sponsor of this podcast.

[Music Playing]

Jim Marous (27:54):

Welcome back to Banking Transformed. Today, I'm joined by Don Permezel, the VP of Data & AI at nCino. We've been exploring the challenges, opportunities and strategies related to the collection and use of data within financial institutions.

Jim Marous (28:10):

So, Don, when we look at this data literacy very significantly across banking departments, how can organizations upskill their employees to be able to use all the insights they're having at their disposable, or at least understand where these insights come from and how to make data have that value we're talking about?

Don Permezel (28:34):

Yeah, great question, Jim. And one very close to my heart, and sort of always has been. Previously, in my career, I've always been obsessed with really just improving the data literacy and data focus within the organizations that I've worked for.

Don Permezel (28:49):

My sort of impression has been that essentially, training isn't enough. So, it's absolutely something that organizations should do, and many of them are, is to take a genuine concerted effort to upskilling sort of product staff in particular, engineering leaders as well, in sort of the skills and toolkits that they need to analyze and understand data, and indeed, solution in a machine learning data and statistical world. Great.

Don Permezel (29:23):

But as I said, it's not enough. I think what I know is that organizations actually need to demand that ... these sorts of people on the product and engineering side really need to actually get their hands as part of their day-to-day work into the data as part of their job, not just training.

Don Permezel (29:46):

They need to be working with the data, they need to be not afraid of the new skills that they need to develop to be able to understand, analyze, clean, and then solution with that data. And look, that takes a cultural shift. But I think that cultural shift is here, but it also takes a bold view from the executive.

Don Permezel (30:05):

As you've said, Jim, our CEO is just sort of incredibly data minded and has not been shy in directing the organization that way. But you've sort of just tailwind a little bit where as sort of the need for these roles to increase their data expertise sort of increases, and there's sort of that training and sort of the procedural component and policy component of what each role does and what each role's expected to do …

Don Permezel (30:34):

Actually, the bar for how much expertise and how "scary" sort of this data analysis and data solutioning is, is coming down due to large language models. So, nowadays, in a code generation world, I know many of our product managers, if not all of them, are in the data platform in natural language sort of inquiring, “Please write me a query that will join this and this and show me what correlates to this effect that I want to achieve for the customer.”

Don Permezel (31:10):

They're doing that in real time. Now, certainly, it's absolutely beneficial for them to have a level of SQL proficiency so that they can see what comes back and make sure it looks broadly correct and all these sort of things. But the bar's going down just as the cultural shift is pushing up that desire to have greater literacy across the organization. And as you said, support from the executive is key.

Don Permezel (31:35):

I think there's probably two more elements. One of which is access to data itself. So, I can tell you one place, especially when you're talking about cultural aspect, there could be other aspects as well, by the way.

Don Permezel (31:48):

One part where I see some institutions get it wrong is by having their default answer of the employee should not have access to that data. As I say this, I know, of course nCino is beyond rigorous when it comes to data … we absolutely understand the privacy and data security requirements of our customer base, and of course, by nature of what we do for a job, we have to take that exceedingly seriously.

Don Permezel (32:21):

Having that said, you can swing much too much the other direction. And I see institutions do that where they really lock down their data as much as possible. That's really locked down. And the sort of requests that you have to raise to get access to data that doesn't threaten anyone's privacy, that doesn't breach any contractual guidelines that the clients themselves would be absolutely delighted with using that data to create more effective products and services. They can take too stricter view.

Don Permezel (32:53):

And that relates me a little bit to a governance point. And another place where I've seen institutions come a cropper in the past is in the governance side around AI and data, where there's sort of this mistake made that minimization of a risk profile, the best way to do that is sort of to pile on more governance, to put on more data restrictions and just pile on as much as this as you can to reduce that risk profile.

Don Permezel (33:21):

And quite often, it doesn't have a huge impact on the risk profile and actually grinds the sort of things to a halt, much like the cultural change that you're asking about. Because how are you going to get an organization that's data-obsessed when the people can't get access to the data? How are you going to develop new AI solutions when the governance process is so overbearing that it just grinds to a halt and nothing comes out the door?

Don Permezel (33:44):

Something which nCino's done very well, and a great credit to our legal and risk teams, is developing both a data access and a governance perspective around AI and data. It's much more like a sniper rifle than it is a [inaudible 00:34:01].

Don Permezel (34:01):

As in, it's absolutely intelligently and soberly looking at the level of risk created to our client base. It's absolutely looking at real security concerns and then saying let's minimize those, but then let's create the ability for us to create value for our customers. And then yeah, so I think that's my answer.

Jim Marous (34:26):

Well, it’s interesting, because the democratization of data is interesting because financial institutions, when I go back to my banking days, there was one area you had to go to get the information. It usually was late, it wasn't clean, and it wasn't even deployable really because I was getting results six months after I did a program as a marketer.

Jim Marous (34:47):

And that becomes really challenging. And the term knowledge is power comes from the, I think it's the 16th century. It's still true, but on the other hand, the democratization of data and knowledge is really where the power lies. And the more you can distribute it across your organization, the more valuable it becomes and the more utilized it becomes.

Jim Marous (35:09):

So, when you look at your organizations, can you share some examples of how an organization has really transformed their culture to be more data driven and what were their strategies to do so?

Don Permezel (35:23):

Yeah, so I mean I can firstly speak from my own organization as nCino, but I can give some examples of some financial institutions that I've worked with in the past.

Don Permezel (35:34):

So, at nCino, we've seen much more ... essentially, it's much like what you said, Jim. In fact, it's basically exactly what you said. We have adopted a strategy of data democratization where those data platforms that I talked about before, and we've spent years and lot of money getting all the pipelines right, all the data security right, all of that sort of thing.

Don Permezel (35:53):

As that have come to the board, we've actually said, guess what, in a responsible way, let's open this up to … obviously, with the right access controls and what have you, all the correct throat clearings. But let's have product managers, when they come in and form a view about the prioritization of their products, let them explain why and show what the data's showing so that the conversation changes from instinct and also feedback, but sort of ad hoc feedback to looking at the numbers and making decisions and prioritizations that way.

Don Permezel (36:25):

And as I said, maybe it's slightly controversial, but I think that as well as the softer cultural element of support from the CEO, executive, sort of other people talking about data and meetings and having it more part of the conversation, I think also a gentle harder push from a policy and procedural standpoint is really valuable too.

Don Permezel (36:47):

Like for instance, saying, "Hey, you want to go and invest in something, you need to actually access the data and figure out how to knit it together such that you're supportive of that respective business case."

Don Permezel (37:00):

What I've also seen is some of the more leading banks who are doing very well, be open to a much more deeper understanding of the processes that underlie their institutions. And not only that – be able to, especially the entrance into the nCino Data Community, which is one of our financial institutions, wants to contribute their processed data into a pool and then receive value back that can be derived from that pool, doing things that wouldn't have been common in banking 10 years ago certainly.

Don Permezel (37:38):

We're able to provide functionality where banks are able to see, here's how my bank works. Like here's my process, here's all the variance, here's what's creating cost, here's what's creating risk, but here's where I look different to other financial institutions that have some similarity to me.

Don Permezel (37:55):

So, talking about this sort of, one, initially, I received some commentary that said, "Look, I'm not sure I want to be more aware of my peers. I know that's the temptation, but I want to really just focus on my customers."

Don Permezel (38:10):

What sort of evolved in those conversations and indeed, prototypes, and all this sort of thing as we're developing is that these more modern banks with a more modern data mindset, were able to sort of say, “Hey, leveraging this data, including this data I'd never had access to before, I'd never had any idea about being able to compare how my bank works to other banks, it actually gives me a better understanding of my own institution and how I create customer value and where I differentiate, and also what's going wrong, how can I fix things”

Don Permezel (38:42):

And so, that much more modern mindset to use data in different ways that have been used before and data that hasn't been used like new data sets, I think is an indicative of change.

Don Permezel (38:54):

The third one I'd say is like when you're deploying artificial intelligence or machine learning solutions, you really need to be able to pinpoint the exact point in the value chain that they're going to create the most value. And then you need to be able to very carefully measure the amount of time and cost release that they're creating.

Don Permezel (39:17):

AI is a world that's rapidly evolving and there's a million bells and whistles left, right, and center and very exciting things, and you can do all sorts of things. But banks need to cut through all of the excitement and get down to exactly how much money am I saving with this technology and exactly, how much growth am I creating with that technology. That's just one on one.

Don Permezel (39:41):

And so, where I see banks have a lot of success, I can think of an example of a large Australian bank where they were previously had a process for essentially analysts extracting numbers from financial statements to do the spreading as part of the credit underwriting on commercial deals. They'd done it that way for many, many, many, many, many, many years.

Don Permezel (40:00):

What they did was they took a trial of nCino's automated spreading functionality, which is a machine learning approach, and then looked at that hard operational data and exactly how much time is exactly how much a user’s spending it each time in the process, and let me do a little bit of a statistics to understand that better.

Don Permezel (40:17):

And this is right from an executive level down, by the way, I had huge support throughout the organization – go and understand that well. And then what they said was, they said, "Look, it's not optional, it's compulsory.” We're using the machine learning technology with all of the change management and sort of slowness to change and all that sort of thing that that implies.

Don Permezel (40:40):

And because they'd sort of done the science to begin with and using that, they decided where to apply the machine learning technology, and then use that as evidence to spread it across the entire bank. They got tremendous results, not only in the numbers by the way, which we could see and help them with, and how your times are getting better, you're saving money.

Don Permezel (40:59):

Actually, the one that sort of just resonated with me a little bit is they sent me an email, which was from one of their bankers who said, "Look, I had this deal. I had to get it in by 5:00 PM on Tuesday, I had 15 financial statements to get through. It would've taken me about seven or eight hours. I got it done in 20 minutes with the machine learning.” So, initially, I was hesitant to do the new thing, but now, I can see the value.” So, yeah, few, little stories of some successful institutions.

Jim Marous (41:31):

Well, it is interesting because as you look at the overall environment we're in right now, we've talked about the deployment and use of data on an outward basis. But I think as you democratize, you're also going to see different parts of the organization being collection zones, because the beauty of generative AI is not the output as much as what I'm going to call the dialogue, where it’s a back and forth – the asking for information to be able to deploy information, the prompting skills that become necessary as you're working just as an individual with generative AI.

Jim Marous (42:05):

But as you look at organizations and the number of touch points that an organization has with the customer or with other departments, we're going to see more and more democratization of the collection of data to be able to deploy. Because those little pieces of data, that knowledge, the looking a little bit deeper into transactions to see where flow of funds are, is really where the real value starts.

Jim Marous (42:30):

And that's an exciting time, it's exciting for a person … not a data person like me, but to know that what's possible. And I think what's going to push the envelope is that consumers are becoming much more aware of what's possible.

[Music Playing]

Jim Marous (42:45):

They're hearing on the news media, they're seeing it in their restaurants, they're seeing it in their grocery stores, they're seen on their ride sharing, on their shopping experiences, and they're going to expect it from their financial institutions who they know have more insight into my daily life than virtually any partner I have in my life, and they're expected.

Jim Marous (43:04):

Don, thank you so much for being on the show, I really enjoyed talking to you and getting a better idea of where data is not going to be a standalone entity anymore. It's the source of value and the holy grail there is, to take it from data collection and deployment to the reality of bringing value to the customer that will bring value back to us. So, thank you so much for your time today.

Don Permezel (43:30):

Thank you, Jim. Been a pleasure.