Everest Group IDP
             PEAK Matrix® 2022  
Indico Named as Major Contender and Star Performer in Everest Group's PEAK Matrix® for Intelligent Document Processing (IDP)
Access the Report

Interview with Automation Center of Excellence expert Brian Anthony

Christopher M. Wells, Ph. D., Indico VP of Research and Development, talks with Brian Anthony, Chief Data Officer at the Municipal Securities Rulemaking Board, in episode 7 of Unstructured Unlocked. Tune in to discover how enterprise data and automation leaders are solving their most complex unstructured data challenges.

Listen to the full podcast here: Unstructured Unlocked episode 7 with Brian Anthony

 

Christopher Wells: Hey, welcome to another episode of Unstructured Unlocked. I’m your host, Chris Wells, VP of R and D at Indico Data, and today I’m really excited to be joined by Brian Anthony, who is Chief Data Officer at the Municipal Securities Rulemaking Board. Brian, how are you today?

Brian Anthony: I’m doing great, Chris. How are you doing?

CW: I’m doing really well, especially since I just got through the name of your organization flawlessly from here on after, to be referred to as MSRB.

BA: Absolutely. So it is a bit of a mouthful and, as, as our organization, as they say, rulemaking’s in our middle name. So, you did, you did fantastic with that.

CW: Excellent. Well, I’ll, I’ll try to stay within the bounds today. <laugh>, this, uh, you, you, uh, you’re more of a data personality than our, our typical guest, which is more of an automation personality. So this is, as someone who is also, uh, you know, a data person at heart, this is a really exciting episode for me. Could you tell us about your career journey, uh, to, to this point, and then what you all do over at the

BA: Sure. So, again, thanks for having me, Chris. Uh, so I will tell you that my career at, at, I started, you know, 30-plus years ago as a software developer. I guess at that time, it was a mainframe developer. , did a lot of cool ball and comics and all that good stuff. , but I have spent most of my time in financial services in some aspects of kind of a technical solution or a data solution. So, again, mentioned that a lot of mainframe coding to begin with. I eventually evolved to some of the client-server and web stuff and then transitioned into data and have never looked back since then, have worked for large financial institutions, specifically in either the securities business or in consumer real estate reporting, and those kinds of aspects as well. So financial services are kind of the heart of what I do, and data is, is a great compliment to that. So in transitioning, <laugh> Municipal Securities rulemaking board is the principal regulator of the municipal bond market. So we’re charged with safeguarding this 4 trillion market. , our mission is to protect issuers and investors simultaneously, and we do that a lot by regulating and writing rules for those that transact and municipal security. So that would be the dealers and the municipal advisors. Data’s a big part that rule-making aspect of this as well.

CW: Excellent. Uh, for the viewers and listeners who don’t know these things, what would you, what would you say are the biggest differences between the muni bond market and, you know, and the sort of the rest of the fixed income market?

BA: So, I think the biggest thing is a couple of things, right? I think the complexity of the offerings for municipal bonds is something that’s a bit of a challenge and how they’re structured. And a lot of that has to do with the tax advantages of municipal bonds. And, so, the structuring of debt that’s associated with that, , definitely makes a difference. Uh, the complexity in the offerings is a little bit different. , you have a lot of sort of tiered bonds or ladder-type bonds that you don’t typically find in the corporate market. And then I think it’s a little bit more complex because the issuers of bonds are, uh, in the corporate market, they are subject to regulatory authority by Fender and S E C specifically. Yeah. And in our market, the regulatory authority is actually to protect issuers, , more so, so it’s a little bit of a nuanced type market.

CW: Yeah, that’s, that’s a huge difference right there. , so financial services is something you and I have in common. I spent far less time than you did there, but, uh, did about a decade. I saw financial services go from being, I would say in a lot of ways data backwards and technology backwards to really starting to take advantage of things like the cloud, even ML technologies, things like that. In your time, going all the way from, uh, global mainframe programming to where you are today, what, what do you, what do you see as, the biggest changes in financial services in terms of tech?

BA: Yeah. I, I think you’re, I think you’re right in how you describe it. I think, tech was, was almost sort of a necessary evil, so to speak. Yeah. , we did it, , we did, we did it because it had to be done to facilitate a problem. What I think we’re seeing more of is the opportunity space with tech and with data and financial services regulators in general, embracing that opportunity space, right? It’s not just a means to an end; it is, in many cases, the value proposition itself. So that’s, that’s really kind of the evolution that I’m seeing more of is really the value proposition with tech and with data.

CW: Yeah. It’s not just taking your medicine anymore, right? 

BA: Absolutely.

CW: Yeah. Talk to me a little bit more about it, you mentioned the regulatory space, which I think a lot of folks who spend time in tech think of as being just, you know, lawyers, yelling at people and getting people in trouble. Talk to me about how regulatory bodies are embracing tech these days.

BA: Absolutely. And, our CEO, Mark Kim described this when we had sort of a career day, where we do have lawyers absolutely within this, but we have, we have data scientists, we have business people, we have industry experts, we have technologists. So it really is incumbent upon us to understand the landscape from our, from our stakeholder’s perspective. And writing the rules is one aspect of that, but how they do business is another aspect of that. And so we try to bring in kinda the right expertise to be able to facilitate transparency in this market, not just purely regulate it, but to facilitate transparency in this market, is probably the best way to describe that.

CW: Facilitate transparency. I like that. Uh, and speaking from personal experience, if you can get lawyers and data scientists to work to well together, you can do a lot of really good stuff. , that’s,

BA: But yes, still, but yes,

CW: It’s getting people to speak the same language.

BA: Absolutely. I mean, that’s really a great point. And that’s where sort of I try to evangelize a bit, Chris, is

CW: Yeah.

BA: I think it’s, I think it’s probably easier for, let me rephrase that. I think as a data person, my goal is to be able to speak more of the business language rather than trying to have the business speak a data language, right? And so that’s why I would challenge our data scientists, our data engineers as a potential skill, right? Is learn to speak the business language. Because I do think if we can bridge those gaps, then there’s a lot more that’s capable for us. So that’s one of the areas I would just challenge my sort of fellow data professionals as well.

CW: Yeah, I, I think that that’s, uh, that’s some good news that should be spread out in the world. , I’ve seen it in my career as well, where it’s both sides need to increase their literacy. Uh, absolutely. But, but really like making the business people learn the data lingo is like, you know, you know, bringing the plumber to your house and him expecting you to know which wrench to use. Like, it’s, it’s crazy, right?

BA: Absolutely.

CW: I think data folks need to really take on a product mindset and say, like, you’re my user. What’s the problem? What’s the opportunity? And let me figure out how to, you know, how to get you a prototype or or get something into production based on your requirements.

BA: Absolutely. Yeah, absolutely.

CW: Excellent point. So let’s transition a little bit to talking specifically about, you talked about sort of tech now not being such a four-letter word, although it’s a four-letter word in financial services. , how have you seen the use of data, say, in the last five to 10 years, uh, change within the organizations that you’ve been a part of?

BA: So I think there’s always been a big t a big sort of tendency in the past that the more data that we collect, the better that we are, right? Yeah. And I think the transition I’m starting to see is more about how we leverage the data that we have in more meaningful ways, right? It’s not about; it’s not about more data necessarily. I mean, I think that’s happening automatically just in the age of technology and smartphones and smart devices and everything else. I think it’s how we leverage the data, and I think that’s, that’s really where we see kind of the opportunity, with all of this is, is just how do we be smarter about how we leverage data, not just consign it for the consing it and storing it.

CW: Yeah. Just taking up space on tape or whatever, <laugh>so that’s interesting to me. The, you know, the change in mindset to how do we make, you know, the most from what we have. What do you think has driven that change in mindset? Is it technology? Is, is it just education? Is it just we’ve been around this much data for so long that we’re starting to think more deeply about it? Where would you point your finger there?

BA: Uh, I think it actually may be kind of twofold. I think it is the fact that we’re inundated with data, and we’re looking for more, , insightful decision-ready, but insightful uses of the data. And then I think the tech is coming along to help deliver that insight faster, quicker, and more meaningful. I think it’s more of a craving for insight, not more data or raw data, right? And I think, , I think it’s becoming, uh, I feel old, but I would say with the younger generation, I think it’s starting to be more accepted, right? If you think about a, if you think about a GPS, right? And you start looking at all of this raw data, you’re not getting all of the step-by-step traffic information of everybody. You’re getting an aggregate state of traffic information that says, it’s busy in this area, it’s gonna take you 15 minutes, it’s gonna take you 20 minutes. That’s the type of information that we need to make decisions on. Yeah. Not every data point that everyone else has collected on your route,

CW: Right?

BA: Yep. So I think that’s the difference.

CW: Yeah. So both the ability and the desire to package the data in the right way are absolute. Have changed. You used a phrase, which I love and wish I could take credit for because it’s so brilliant, uh, but you used the phrase decision-ready? Yes. Talk to me about what decision-ready data looks like.

BA: So the fed’s decision ready decision-useful, I wish I could take credit for it, but it’s, I heard a little speech with, uh, secretary, uh, with, excuse me, the chair of the s e uh, chair Gensler. But it really is about the investor community, the business users, and the stakeholder being inundated with data. So how do we present them with meaningful insights that can aid, not necessarily replace their decision-making process, but aid their decision making process? Because we don’t want to take the intuition, the savvy, the experience, the human experience out of this. We wanna complement that. So I don’t suggest that data gonna automate all of our decision-making. I think it should be a tool that’s part of the decision-making. I think the decision-ready, decision-useful data, means that you’re getting the insights you need when you make the decision. Kinda going back to that GPS example, if you think about it, if I’m already in the middle of traffic and that’s when I get the notice that there’s a traffic jam <laugh>, right? Or, or that, you know, that’s not really that helpful for me at that time.

CW: No.

BA: So that’s how I see sort of the decision radio decision, useful data. And again,, I’d love to take credit for that one, but that’s a chair getler. 

CW: Well, here’s the, here’s the chair. Getler. It’s great, it’s a great phrase. I think, uh, it’s interesting. I think financial services, especially like the back office in your investment banks, organizations like yourselves. Then, you know, the larger insurers, as I mentioned there, there’s been sort of a renaissance in the last few years of thinking about data in this way. And it’s interesting, like, you know, healthcare, especially in the radiology space, got there so much earlier in the sense that they realized it’s really hard to get an AI to read, you know, radiology reports, x-rays, CAT scans, with the, with the right amount of precision. But that read plus, uh, the human expert coming in and sort of over the top making sure it’s correct, like massive improvements and, you know, throughput and accuracy and all of that stuff. So, you know, all good things come to those who wait, I guess, uh, it’s, it’s our turn in financial services to get there.

BA: Absolutely. And I think, you know, the healthcare services and, and a lot of those types of tech firms or, or data firms have helped pave the way, and in a lot of cases, it’s been a little bit more of an expensive endeavor because they were kind of first to lead that effort. And so when you start, you know, go back, going back to your question about, you know, it being more prevalent now in financial services, just a more conservative risk of risk-averse type environment. Yeah. Now being able to leverage more proven cases, models, those kinds of things. I think that’s the other thing that’s sort of driving adoption at this point in time.

CW: Yeah, I think you’re right. The technologies have gotten better. , and a lot of those people who failed early in some other spaces they’ve made their way out into organizations, absolutely.

Just a better collective understanding of what’s possible. Yeah. Great point. Brian. , how do you decide with the massive, you know, massive amounts of data, both structured and unstructured, that are coming your way, how do you decide with all of that data what the right data is to work on at a given moment? Like, you could, you could choose any number of projects I’m sure to be working on, and you probably are working on a dozen right now, but like, how do you prioritize, you know, what this data is gonna provide the insights we need, let’s spend time refining it.

BA: So that’s a great question, Chris. And I wish I had a magic eight-ball to get all the projects right all the time, right? Yeah. But I think part of it really, it goes back to the stakeholder of value. And I think, , there’s a little bit of experimentation, there are some knowns, right? That you can work on that has a high value, have high values, excuse me. There’s a little bit of experimentation in some cases. And I think especially with those, right, partnering with the right stakeholders, learning early, , the whole MVP concept, really learning early if this is viable or not, and whether or not it’s going to have a desired impact on stakeholders. Because at the end of the day, it can be a cool data project for me, but it may have zero stakeholder values. So yeah, really, it’s going to be the stakeholders partnering with you along the way that’s going to sort of reinforce or help reinforce if you’re moving in the right direction. And so, again, that’s why I go back to; it’s incumbent upon me, my data professionals, my tech professional peers, and my colleagues to learn and speak the language of the business a little bit more.

CW: Yeah. You’re, you’re telling me, uh, the F1 score isn’t a, isn’t a good enough answer for the business

BA: You’d be surprised. Yes. I, I know that I could explain the F1 of lawyers and business partners that I work with.

CW: Yeah. Nor should you. As an R and D lead, one of, one of the mantras I repeat all the time is that engineering begins with certainty, right? You know what you need to build; you go build it. Research ends with certainty, right? We’ve done the experiments, and we’ve made conclusions. How do you set yourself up as you’re doing this experimentation and prototyping? How do you, how do you design good experiments? Like where you’re not gonna get to the end of just shrug, because you can’t tell whether it was a success or failure

BA: <laugh>. So, uh, the, the key part to that, right? And, so first of all, I love that mantra, by the way, and, just kinda before getting into that, I think part of what setting yourself up for good resources, for good researching is also, predicated on having those good engineering practices, right? And having done that work ahead of time. So all of the basic information, right? The nice roles and columns and structured data, let’s get that cleaned up as much as possible so that we don’t spend an extraordinary amount of time every time we do a research project going back through that. Yeah. So, I love that engineering side of it and getting that squared away as much as possible so that the research is not as expensive each time you do it, the first part of it.

But I think it goes back to defining your success criteria, right? Or defining your hypothesis. , getting back to sort of the scientific notation side of this, Yeah. What am I trying to prove? And if I’m going through this research process and I’m disproving my hypothesis, that’s not necessarily a bad thing, but I know when it’s time to, I know when it’s time to, to say, you know what, this, this wasn’t what we were looking for and, and maybe we punt now and, and look at something else. But I really think it’s a clear hypothesis. I think it is. It is also some defining, business objectives from this as well. And if we’re trending in that direction, the cost benefit analysis is still within our ranges, and let’s keep going if it’s not, and let’s cut our losses and, uh, look at other opportunities.

CW: Yeah. So clear design up front, knowing what success looks like, and then holding yourself accountable frequently, uh, and honestly throughout the process, right? Is that, is that what I’m hearing?

BA: Absolutely. And I think honestly, it, the honestly part is a, is a good aspect of this as well, because we can tell ourselves that we’re, oh yeah, we’re, you know, we’re still on track. We’re still doing well. , so I, I, I think that’s probably one of the hardest part about this is cause you’re so vested in some of these things that Yeah. , that it’s hard that it takes a mindset to know that, you know what, this, with an experiment, it’s okay to let it go. Yeah.

CW: Yeah. Everyone prefers a yes answer to can’t we do this? Right? But I think back to when I was an academic, I had a shelf with my research on it, and there was a pile of things that worked, and then there was a pile of things that didn’t work, and it was like 10 to one. And, uh, it was a good lesson to learn early in my career. , alright. We’ve been floating out in the abstract a little bit. Let’s come back to the concrete world. You’ve been at RB now for about three and a half years, I think, if I remember correctly. Yes. Give yourself a chance to brag about your work and your team a little bit. What are you most proud of, uh, in the, you know, that you’ve accomplished in the time you’ve been there?

BA: So what I’m most proud of and what I think I’m most proud of about the team, frankly, has, has nothing to do with technology or data, actually. I think it has to do with the culture around data that we’re helping to establish, right? And the importance around data that we’re helping to establish. , I think the, our organization has historically been and, and with good reason. , our intent has been to collect the information and distribute it as quickly as possible. And our mindset has been hands-off, this isn’t our data or our responsibility. And, there were good reasons for that. And so to begin to change that mindset that says we have a responsibility for the quality of data, that we have a responsibility for improving our stakeholder experience with this and creating those opportunities for that to be starting to be embraced by the organization.

CW: Yeah.

BA: I’m very proud of our team for, , being champions for that, right? And starting to help distill that across the organization, , is huge for us. And then we get to some of the projects that we’re doing that are fine and interesting and sort of support that, but that’s how I would sort of, uh, classify.

CW: No,, I love that answer. Uh, it’s, it’s relatively easy to change technologies and, you know, do good project management and even product is harder, but changing the culture of an organization is like a monumental task.

BA: , but we’re not there yet, but that’s, we’re, we’re moving in that direction.

CW: Yeah. But you got, you got a grasp movement. Yeah, no, that’s, that’s great. Uh, talk to me; this is something that I think our audience could, uh, would love to hear. Talk to me about what a healthy data culture looks like.

BA: So, it’s a great question. I, I think, for us it’s an appreciation for, the purpose of data. I also think it’s appreciation for, how data can be used throughout your business processes, throughout your stakeholder processes. , and it, and beginning to embrace that. And, and when we can get to the point that as an organization that we’re thinking that we’re incorporating that data or that insight into our decision making processes intuitively, right? That to me begins to be a healthy, culture around data, right? And so kinda going back to the conversation about financial services and collecting data and collecting data and collecting data to what end Yeah. Right. To what end, right. And so to begin to realize what it is that my business, my stakeholders need from all that information that we’re collecting, and provide it real time or near real time, and begin incorporate that into our, even our own internal processes, right? That’s a healthy culture in, in my mind, a healthy data culture in my mind. I think, I think business leaders offer to rate on intuition. If we can complement that to support their intuition, or what’s even more interesting is for a business person to change their opinion because of the data, right? Right. Yeah. That’s another aspect of just sort of conveying a healthy business culture, but a healthy data culture. But, uh, yeah. Yeah. So that’s, that’s key. But

CW: Yeah. And, and, uh, you sort of made this point implicitly, but healthy data culture is gonna drive healthy data practices, Absolutely. Both upstream and downstream, right?

BA: Absolutely.

CW: Yep. , it’s interesting, you know, you’re, you’ve been in and around the asset management world, and I can’t tell you how many times I’ve seen folks also who have been in and around the asset management, you know, world that don’t think of their data as an asset. , I worked with one firm a while back where, uh, you know, their grounds staff knew where every office chair in the building was and how old it was, and, you know, when it was due for, you know, sort of being cycled. But they couldn’t tell me, you know, how many of a certain type of bond docent they had, nor did they know where to go look to answer that question. It’s like, I, this is an asset, you’re, you’re not managing it well. , and that, you know, and it led to bad data practices cuz they didn’t have a culture of valuing what they had.

BA: Absolutely. And I think what you just said is, is key, right? Is I think recognizing that data is an asset, right? And managing it accordingly is another indication of, of a healthy data culture as well.

CW: So Yeah, absolutely. All right. Well, let’s, let’s turn our attention to the elephant in the room, which is unstructured data. <laugh>. Give me the, uh, Brian Anthony, chief data officer definition of unstructured data.

BA: Data. So reams of information, free form information or textual information that it takes more work to apply context to that, that really is unstructured data, right? I mean, if I can apply basic engineering technologies to aggregate something to derive some prediction out of. But to be able to get to the context that’s being expressed, right? The sentiment that’s being expressed, and sentences, paragraphs, there is so much information, so much valuable information that is still being expressed that way. , and so how do we get to that decision, useful decision ready information out of this, these, uh, text, in paragraphs, sentences, those kinds of pieces of

CW: Data. Yeah, I like that answer. I always ask this question, and I always get a slightly different answer, and you’re the first person to really highlight how important context is for understanding unstructured data. Talk to me a little bit about how that plays out in what y’all do with, , you know, bond docents, for example.

BA: Absolutely. And it’s significant for us because the phrase that was used when I was joining the organization is that we’re sitting on a treasure trove of information. And that could not have been a more true statement. And so let me describe for, for the listeners sort to put that in context is that we have, we have several rules where we collect data. A lot of it’s unstructured data. One of the rules that, of data that we collect is referred to, as an official statement or a bond offering docent. Okay? And this docent is very much like a prospectus, but it’s also very much like the reams of information that you would supply if you were doing a home loan is telling you about the health, you know, health and welfare of the, the person who’s issuing the data, the financial information, the purpose of the bond, how it’s structured.

It’s, uh, oftentimes a two or 300 page legal docent with very relevant information about the bond. That’s one type of docent that we get for every bond that’s issued in the muni bond. In the muni space. The SCC has a rule, where there’s, it’s referred to as a continuing disclosure rule, where as long as that issuer is participating in the bond market, they have an obligation to supply, certain information to, again, convey the health of that issuer, throughout the life of that, just making sure that that information’s transparent for the issuer. All of that information comes in an unstructured docent, uh, financials, uh, bond calls, the faults or, or, or, uh, any kind of material information that might affect, an investor’s decision. All of an information comes of unstructured data through our systems, right? So we are sitting on a literal treasure trove of data, right? And a treasure trove of insights. And, our role has been traditionally to get that information back out into the marketplace, but also in an unstructured format, right? So we collected and unstructured, and then we immediately disseminated back out and unstructured. So now you got tens of thousands of people or organizations who are all doing the same thing, which is trying to derive insights to understand the health of the bond market.

CW: Yep.

BA: So if there are opportunities for us to begin to extract and disseminate pieces of that so that we’re providing insight, , then everyone doesn’t have to be doing the same thing. And actually, there’s probably opportunities for us to collaborate more so that we’re doing different aspects of transparency. , so it would be interesting to see kind of how we continue to explore that. But that’s, that’s kind of our use case around structured data. And yeah, I checked Chris, and since January of Twitter, I believe that we’ve received somewhere around 300,000, PDF docents, which is somewhere around 28 or 29 million pages of text.

CW:

Wow. <laugh>. Yeah. That’s a treasure that counts as a treasure trove.

BA: That is definitely a treasure trove.

CW: So maybe even the mother load. Yeah.

BA: Any insight or any ability to unlock that content is really key for us. And, really kind of the first aspect in unlocking it was just to make it searchable, right? Mm-hmm. <affirmative> and in a, , convenient way. So searching within the content, searching across the content. , so that was really kind of our first opportunity, and we’re proud to have released kind of our ML labs platform, which enables searching. , but, but you and I both know it could go a lot further than searching, but, but even that was such a huge milestone for us as well.

CW: Yeah. No, I, look, you’ve got 29 million pages. , the ability to be able to find things in without actually literally reading through everyone is a huge step, right? Like exponentially decreases. , there’s a lot in what you said, so let me try to, let me try to pull a few threads. , one, uh, I think it’s really interesting, you know, you talked about context being important and you talked about like where you are in the docent as part of that context, but you also highlighted an element of this, which is like, when it comes to you, whom it comes from, and, uh, you know, the relationships among the different docents, because there’s some sort of central entity, , or bond offering that’s common. All of that is part of the context, right? And what makes unstructured even harder is maintaining those relationships.

BA: Absolutely. And those are some of the challenges, right? I mean, it is, it is the nature of the docent, the metadata associated with those docents, , and in the context, the meaning that they’re conveying as well, that, that all have significance. And how did you begin to digest, organize, and disseminate that significance, not just disseminate the docents themselves, right?

CW: Yeah. That’s interesting. I’m sure a lot of folks have worked with or, or seen, you know, sentiment analysis on language, but in the financial markets, the language is all very sterile legal. And so, you know, getting at things like, how significant is this, to me, that’s a really tough, uh, ML problem. , and, uh, it sounds like you all have the right data to be able to solve it,

BA: <laugh> Well, we have the right data as the challenge. Yeah. , solving it is definitely, definitely the challenge with step

CW: Step

BA: Two. And, and it’s funny because you talk about, so earlier in the conversation you talk about which problem are we trying to solve and how do we decide that? And, and that’s an even greater challenge in deciding where we go with unstructured data, because again, there’s so much context that’s being conveyed, right? There’s, there’s just, just generally there are opportunities that are being conveyed in, in this data. There are risks that are being conveyed and this data, right? And then, and then there’s just the status quo that’s being, you know, and so how do we begin to dissect which problem is the most important for the industry for us to invest resources in with the industry? Is it esg, environmental social governance? Is it cyber risk? Is it other risks that

CW: Yeah, all of a sudden there’s a pandemic that there hadn’t done before, right.

BA: Pandemic, right? Absolutely. So, and that was one area that, again, that we started taking advantage of the information that was being provided within these docents is, is even conveying the covid impacts to municipalities. But so which problem are we trying to solve? And when and where are we putting our resources? And where, where’s the industry put against resources? There’s no easy answer for, for those kinds of things, which is, you know, why it’s important to partner with stakeholders, to, to know that you’re succeeding or know that you’re failing and pivot, you know, quickly as well.

CW: Yeah. Again, you’re bringing back that product mindset, which I love. I wish every data person had a product mindset. , uh, another thread I wanted to pull, you talked about how there, you know, there really tens of thousands of stakeholders and before you got involved and you started to, as an organization, try to transform the level of insight that you were providing. , it’s really a situation where all, you know, tens of thousands of those stakeholders have to look at all of the same stuff, right? And so if your goal is bringing transparency, , then, you know, making the easy stuff obvious to everybody, certainly seems like it checks a big part of that box to me.

BA: Yeah. And actually, I mean, that’s really sort of historically the role that our organization has played. It’s evolved in our history, it evolved from rule writing to ML, which is our flagship, uh, sort of transparency system. So we started trying to give context and more transparency to the market. And so really this is, in my mind, kind of the next evolution of that. , which is not only transparency in terms of raw data that’s being provided, but transparency in terms of insight that’s being provided, and how do you, how do you maximize the resources that are available to this marketplace, in terms of let’s solve for or at least partner with, with, with appropriate firms to kinda solve some of the more common use cases. Yeah. As a regulator, we’re not gonna go out and try to solve the most obscure use cases, right? Let’s try to solve some of the more common use cases so that others can solve some of the more obscure type use cases and value add. So it’s, it’s a great way to, it’s a great opportunity to partner with others in this industry, for solutions as well. And, and so we, we kinda do some of the more common aspects of this and let others leverage that, and then let others invest, you know, more private, private firms invest in sort of the, uh, more nuanced type things as well.

CW: Yeah, yeah. No,that’s a great framing. , people often ask me, uh, why should I care about unstructured data? And I, we, you, and I have talked a lot about the optimistic answer to that, which is there’s just a ton of opportunity, right? , for streamlining processes, automation, just insight, all of that. There’s also the pessimistic answer, which is you’re sitting on that treasure, that treasure trove might actually be a landmine depending on what’s in it. So I, we haven’t talked about that at all, but do you see, you know, if you stop doing anything with unstructured data, data today, like maintain status quo, like what are the risks that you see for the

BA: Organization? I think the risk of very similar to the opportunities, right, is that if you really want to make better decisions, then you have to walk into these decisions with your eyes open, whether that’s great news or whether that’s not so great news, right? Yep. And so, doing nothing, I don’t have the context of whether it’s great news or not-so-great news. I don’t have the opportunity to pivot early. I don’t have the opportunity to see or contextualize this risk, , and oftentimes until it’s too late until it hits me sort of immediately. So I think it’s not a question of whether or not it’s conveying good insights or bad insights. It’s conveying insights, right? Yeah. You, as the business leader, as a stakeholder, make your decisions based on that insight, right? So that’s more the way I see it, is that, so take the, it’s good news or bad news out of the equation at this stage and just make it news, right? And then

CW: Yeah. That’s interesting. Uh, you know, people talk about data driven versus data influenced decision making. You’re, you’re definitely leaning on the data influence side, which, I like a lot personally. And on the other hand, I really, I really like the framing, which is you’re not saying like, just because we have this data, you’re gonna make better decisions. You could still make bad decisions, but you’ll have a deeper understanding of whether you did or not and why. Absolutely. And that, that’s huge.

BA: Absolutely. So, and, and, and one of the great aspects about that is if I thought the decision, the data was reflecting one decision, I went in another direction, at least I have something to evaluate against.

CW: That’s right.

BA: Yep. So I think kind of the other side of this is the measurement of your decision making process, right? That has to be incorporated into, into this process. But, this is why I love the data space so much, right? <laugh>, it’s just the opportunities with it. And, and it really is about the business, about the stakeholder, right? Nothing in my in my past experience tells me that anything can influence the business of the stakeholder more than the right data in the hands of the right business people.

CW: Yeah. Yep. You, you, you talked about business leaders being intuition driven, and I, I think the best business leaders are intuition driven and are willing to, uh, update that intuition based on the data that they’ve seen, right. And the decisions they’ve made.

BA: Absolutely.

CW: Well, this is great. Time is flying along. Talk to me a little bit, what, what do you see as the, uh, as you think about the tech space for data, what do you see as like the next frontier then our, the next horizon for, data technologies?

BA: Yeah. So for me personally, I definitely want to see more continued growth in the natural language processing space. The ability to contextualize information easier. I think one of the things I shared with you before, Chris, thinks especially for an organization our size, the ability to leverage existing models, without a huge investment to, to lean on some of the, uh, some of the infrastructures that others have created so that we can take advantage of that without significant investments and being, you know, uh, sort of responsible stewards, you know, with, with what we’re trusted with, I think is huge. Because I would tell you that I don’t believe five years ago that we could be investing in this, or because I don’t think as an organization it would’ve been as fiscally responsible for us five years ago, given where the technology was, or 10 years ago, where the technology was, nevermind, where the heads of, you know, our business leaders, our stakeholders were, I think just technically speaking, it would not have been practical. Yeah, as the technology, the infrastructure, the ability to leverage and reuse, , models and, and infrastructure and framework continues to grow, I think it becomes, I think that moves not only us as a regulator forward, I think it also moves small businesses forward as well, right. To be able to leverage that capability. So as much as it is about specific instances of our AI machine learning, it’s more about the infrastructure as well. Yeah.

CW: Accessibility, right? Absolutely. These things. Yeah. That’s interesting. We’ve talked about this before. I am fascinated by the, uh, AI art generation stuff that’s out there. And it, it’s all very silly <laugh> for the most part. There are a lot of interesting legal and ethical dilemmas that it raises, but it makes me really excited in that, you know, I’ve had, you know, natural language processing in my phone in my pocket for years now, but I don’t really use it that much. Mm-hmm. <affirmative>. , and most people just sort of ignore it. But right now, you know, right now with the AI art stuff, Dolly, stable diffusion, you’ve got regular folks just interacting with ai, every day, and, and finding use from it, and whole new industries are gonna pop up around this.

BA: Oh, I completely agree.

CW: And, I think the same’s gonna be true in the enterprise. Like as, as our culture generally becomes more aware of what’s possible with AI and what isn’t, then that’ll infiltrate the enterprise and, you know, government organizations as well.

BA: I completely agree. And I, I, and I think that’s the that’s sort of the bar that I’m talking about, right? Is that, is that for organizations to be able to find useful ways to even leverage that, the art generation side of this, whether it’s, whether it’s enhancing a, a PowerPoint with some art. Yeah. Because you, you can’t find what you’re looking for in the

CW: Study. Cause I have no imagination. Yeah, exactly.

BA: So a lot of those kinds of things, I think, again, lowers the bar for those advancements and technologies to be incorporated into, into the mainstream. Yeah. Right. So I think that’s what’s, uh, really, really exciting about this space.

CW: That’s great. , okay. This is sort of the other side of that question. Sure, you have spent a career trying to communicate to business leaders about data and what’s possible with data and the right way to do it the wrong way; what do you hope, uh, you know, the MBA student of the future is learning about data and what’s possible with data?

BA: So that’s a great question. I just, so I have a daughter who’s not an MBA student. She’s actually an undergrad student in animal science.

She’s in her senior year. And one of the things that she talks to me about was her use of python and research, right? So she has no interest in technology, but she has to use technology in data languages to do her research and analyze and get results. And so what I would say to students and to the MBA students is it is an enabler, right? For whatever it is that you’re doing. And even if you’re not the one who’s actually doing it, just recognizing that it is an enabler and either learning a little bit more about it or partnering with others to leverage it will have a great impact on what you’re able to do and your level of influence as well. So yeah, that’s what I would say to the MBA

CW: Students. All right. Great answer. I love it. Continue the tradition of not treating tech and data like necessary evils. Keep that going.

BA: That’s it.

CW: Yeah, let’s see. How should we wrap this up? I have a couple of closing questions we choose from, , let me go with this old chestnut. What did I forget to ask you that I should have asked you, Brian?

BA: The importance of what I do and the importance of our organization how would I sort of classify what’s next for us or what’s next within this? And as I think through that, because I, I, I do think that we try to think with the product mindset, I try to think with, leveraging data and tools to do this. I guess where I would go with this is that one of our strategic pillars for our organization is on public trust. Mm-hmm. <affirmative>. And I think everything that we do as a regulatory in, as a regulatory entity has to continue to instill trust and more trust and more trust.

CW: Yeah.

BA: And trust is really fact-based, right? So how can we continue to reiterate the facts in this market, the facts of us as an organization, the fact of us in this industry? And that’s the kind of things that, that I wanna sort of help drive forward for this organization, this industry is the facts, right? Yeah. And so, , that’s where I would sort of close. All

CW: Right. That’s great question and great answer. Didn’t even need me. I should have just turned my camera off and let you talk.

BA: Oh, no, this is fun. I’ve enjoyed a great deal talking to you, Chris, both today and in past conversations.

CW: Well. So, yeah. Yeah. This was fun. Well, I’ll close it out here. You’ve, uh, you’ve been listening to or watching another episode of Unstructured Unlocked. My eminent guest today has been Brian Anthony, the Chief Data Officer of Brian. Thanks so much,

BA: Chris. Thank you for having me. I really enjoy this conversation.

CW:

Okay. Best of luck finding all those facts.

 

Check out the full Unstructured Unlocked podcast on your favorite platform, including:

Subscribe to our LinkedIn newsletter.

Get started with Indico

Interactive demo

Transform your own unstructured documents with our OOTB models

Live Demo

Explore firsthand the value the Indico Platform delivers

Talk with us

Discuss how the Indico Platform can help you tackle your unstructured data problems

Resources

Blog

Gain insights from experts in automation, data, machine learning, and digital transformation.

Unstructured Data Explained

Answers to the most complex questions in unstructured data.

CTO Corner

An accumulation of content straight from our co-founder and CTO.

Unstructured Unlocked

Enterprise leaders discuss how to unlock value from unstructured data.