Indico Data receives top position in Everest Group's Intelligent Document Processing (IDP) Insurance PEAK Matrix® 2024
Read More
  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

Unstructured Unlocked episode 16 with Stacey Brown

Watch Christopher M. Wells, Ph. D., Indico VP of Research and Development, and Michelle Gouveia, VP at Sandbox Insurtech Ventures, in episode 16 of Unstructured Unlocked with guest Stacey Brown, MBA, InsurTech Hartford. Join us as we discuss how fast AND accurate document intake enables them to make data-driven decisions, increase capacity, and drive top line revenue.

Listen to the full podcast here: Unstructured Unlocked episode 16 with Stacey Brown


Christopher Wells: Hi, and welcome to another episode of Unstructured Unlocked. My name is Chris Wells, VP of r and d at Indico Data, and I’m joined today by my co-host, Michelle Govea. Michelle.

Michelle Gouviea: Hey, Chris. Today, We are pleased to have Stacy Brown, founder and president of Insurtech Hartford, as our special guest. Stacy, welcome. Thanks so much for taking the time to speak with us today.

Stacey Brown: Thanks, Michelle. Thank you, Chris. Absolutely. Happy to be here.

MG: Wonderful. so, Stacy, you and I have known each other for a while from the conference circuit and InsureTech Hartford specifically. But, for our listeners, do you mind just giving us a little bit of your background and experience?

SB: Sure. definitely. I think many people have gotten to know me over the years through the work with InsureTech Hartford, but for those that might be new, you know, Stacy Brown, I’ve been on the carrier side of the insurance technology industry for 20-plus years. I’ve worked for companies such as Travelers and X xl, and I’ve had the opportunity to do everything there is to do in the technology space over the years, from, you know, building policy admin systems by hand. When we used to do that, and back in the day before, you know, big package solutions, even AI and ml-related projects, came along. I recently left my role at Excel as head of innovation delivery, where I was, you know, overseeing a lot of different initiatives at the company. So, I also have a pretty broad background in innovation.

CW: Yeah, innovation in the insurance space is something we talk

SB: About in the engineering space. Yeah.

CW: Yeah. That’s exciting.

MG: Awesome. Well, well, Stacy, I would love to do with those two. Two have kind of the technology background, but also the lens and the focus on innovation from my own experience within the insurance carrier and just what I see in my role today, there’s, there’s, there are synergies there. Still, there are sometimes some butting heads between the innovation projects teens want to get off the ground and either the technology budget or the in-house capabilities from a carrier perspective. How did? How did your role? How did you marry those two, and what were some of the challenges or successes you might have had within the carrier pushing most two agendas forward?

SB: Well, it’s always a struggle. I heard somebody say once, so this wasn’t me, but I’ve been re. I’ve repeated it many times because it’s so true that when it comes to insurance companies and technology, they basically can do anything they want. They just can’t do everything they want, right? There are a lot of great problems, a lot of great problems to solve. A lot of every single one of them is, is, is wonderful in its own right, but every organization, organization struggles mainly with two big things, right? Budget and capacity. Only so many people in an organization have the skills necessary to run these projects. And those skills are usually in demand across multiple projects.

And so even when something makes a lot of sense financially, right? For, if you can get past the budget hurdle, you sometimes run into the, now how do we execute this hurdle? And there’s always somebody out there, a provider, consulting firm, or whatever, willing to come in and step in and wear whatever hat you want on your team. But it’s not the same as having your experts engaged, right? So, the budget itself is usually the first hurdle. And, you know, every company’s in a little bit of a different spot, but most insurance companies are, you know, they managed their budget tightly. They’re, they’re trying to meet certain combined ratio targets. And so its spending is typically an allocation for the year, right?

So they sit down and think, oh, we’re going to spend, whether it’s 10 million or 150 million or whatever the budget for the company is going to be on, on technology next year. So when you’ve got a fixed, a fixed capacity and a fixed budget, it leaves you really with prioritization. And, you know, Michelle, you asked the question about, you know, how do you overcome them, the challenges of that, quite honestly, after 20-something years of doing it, you just don’t sometimes because mm-hmm. <Affirmative>, you know, on top of, if you can, make a solid business case. And then, if you could even make the case that you have the expertise and capacity to do it, sometimes other things are behind your control, like I don’t want to say politics. Still, you know, if there’s a, if there’s a, a group in the organization that’s making a lot of money for the company, their voice is going to weigh more than somebody else’s voice, right? So, you wind up having to balance those internal things as well. They come into the factor when it comes down to prioritization. So I don’t know, I hope I kind of went on a long rant about that, but hopefully, hopefully that was insightful.

CW: It, it’s helpful. Background color. You raised the notion of combined ratios, and one of the places where we know folks are focusing right now is trying to fix up those ratios by focusing on, you know, obviously, underwriting and claims. And then we’ve heard a lot lately about the intake process being a focus, especially for technology spend. So what are, what are people doing in those areas today? And what would they like to be doing if they, if they had the budget, had everything they needed?

SB: You know, the, the, the intake process, I guess there are really two main points of, of intake for as I, as I’m thinking of it, and it’s, I think is what you mean. One would be, you know, during the underwriting process, the intake of a new submission, something for an underwriter to look at and try to come up with some kind of price quote that hopefully leads to a bind, right? Yep. the other side of the coin is on the claim side, right? First notice of loss. And, you know, there’s been a lot of, over the years, a lot of different innovations in that space of people trying to come up with apps and, and, and, and web portals and, and things like that that help make claims reporting easier. But still, the fact of the matter is that every day, the first notice of loss insurance carriers are received is a, is letter a from a law firm in the mail, right?

<Laugh>, like it happens every day. So no matter how much structured an approach and user-friendly user journey type of tech and, and approaches we take, you know, there’s still, there’s still unstructured initiation of, of, of the claims process. I don’t know that that’ll ever go away either, right? Like, if somebody hires a law firm because they don’t want to talk to their insurance company first, well, the law firm’s not going to log on to your, to your company portal to submit an F n LOL F N O L request, right? They’re going to write that 12-page, you know, document and stick, stick it in the mail, send it, certified return receipt, and all that stuff, right? That’s fillable

CW: Hours right there.

SB: It, it is. Yeah. <laugh>,

MG: We, we’ve talked a lot about that, Stacy. Actually, that intake can, can look different depending on your line of business. Your, your, the company that you’re working for, right? From various channels. And so to your point, I agree there’s never gonna be a single way that all documentation comes into the carry, either on the underwriting side or on the claim side that what, what we’ve talked about a lot is having the right mechanisms in place to receive it regardless of the channel that it comes in. So from an automation standpoint, what have you seen in your roles in terms of you mentioned it, right, the unstructured data intake, but even getting your systems prepared to then digest that data in a way that’s, that’s usable in the downstream systems?

SB: Yeah, I mean, it usually starts the most common way that these, these workflows starts tends to be through email inboxes, right? So you might get like a call center or a mail center that takes mail and scans it in, and then sends it to a mailbox, and then you get, you might get brokers and agents submitting you know, claims that they’re, they’re hearing into an inbox or, or or even on the, on the policy submission side, right? Email is just like, tends to be the number one place that is, it’s probably the most common place that the process starts from, you know, there could be other fancier things like API based you know, integrations and queues and things of that sort. But really it’s, it’s seems to be inbox. And then you know, there’s still a good amount of people managing those inboxes, right?

So over the last, you know, 15 years as outsourcing and offshoring of, of this type of work became popular the way to save money was just to move the workload to a lowest cost labor market. And you know, that still happens, but what’s happened as a result of that is it’s actually made the business case for automation even more challenging, right? Because if you’re paying somebody, you know, $8 an hour and they’re, and they’re able to process, you know, you know, 30 mail e emails an hour, right? Like, that’s, that’s a, that’s a lot of productivity for a pretty low cost. And and unless you’re getting like a million of these things in a day, it’s really hard to get the scale to make an investment, right? So that’s like one challenge point right there, right? But once, once, so when it comes to automating the process, the somehow or other, you have to be able to break down that document, identify what type of document it is a claims document or a submission?

Or is it a le is it a bill that should be going to the procurement team or whatever, right? There’s, there’s all kinds of, is all kinds of things that those letters can be. But then let’s say you find that, oh, this looks like a claim from a potential claim from lawyer. You know, you need to go through there. You need to, to extract, the entities, right? You need to know, like, try to figure out who is the, the claimant, who the policyholder is, and what’s the law firm if there is one even. So it, it, it gets, it gets pretty complex to, to, to automate that process. And over the last several years, I’ve seen a number of initiatives with different technologies trying to figure out, you know, how do we, how do we classify the documents?

How do we identify the, the data sets, you know, taking first, you know, OCRing the data putting it into sentences, then using sentence fragmenting technologies to identify the, subjects? And ultimately you know, trying to extract the data, it’s just been a real slog. So what’s been happening is there’s been this trend to go towards specialization when it comes to the AI. The natural language processing technology is required to do that, that kind of extraction. So in, you know, there are companies out there, you know, that when it comes to the technology and the, the frameworks around ai, whether their ML-based or, or rules-based, you know, there’s, there’s, there are only so many patterns, right? And so everyone at some level of ab abstraction is doing the same thing, but with where people are, are choose, are choosing to focus and go really deep into just like, you know, claim first, a notice of loss for, for, for homeowners policies, right?

Like that, that is increasing success because you start getting enough volume in, in order to to do the training. And, and, and actually that’s like a whole nother challenge, right? So a check another challenge is just making sure that you have the volumes of data to train these models. And so where there are providers out there that are able to use their learnings across, you know, multiple carriers you know, that is a way of, of, of training models better and, and faster. And that’s, you know, from there, once the data’s extracted, it needs to get from that unstructured format to the structured format. And that usually means that whatever the downstream systems are, it’s usually you’re starting out with your policy admin system. And, and so most policy admin systems today, IP groups insurance companies have invested in their IT to, to put APIs out in front of them.

 Everything that’s out of the box today comes with out-of-the-box APIs. So the good thing is, I think we’ve gotten to the point as an industry where, you know, I don’t have an exact figure, but I’d say it’s a very high percentage of the solutions out there are, are API enabled, right? So, so if you can get good at extracting this data, now you need to somehow map that data to the structured message format of these APIs, right? And then once we can push what so that in and of itself could require some magic in terms of AI and, and, and, and things of that sort. So

CW: Sorry, I’m going to jump in because my list of follow-up questions is now almost more than I can keep in my head. There’s a lot of good,

SB: Sorry. No, it’s okay. Go, go for it, though.

CW: The first one is, you know, given you, you mentioned that volumes have to be really high to make a case to cut off the outsourcing and bring it back in-house with automation. So does the return on investment argument shift to looking like repeatability in the process and therefore cleaner structured data? Or is there some other tact that you see folks taking to make that automation case nowadays?

SB: Definitely a quality is the thing. Cuz when, what one of the downsides of, of using the lowcost labor approach is quality. You know, because people type and make mistakes when typing. And you know what, computers make mistakes when they’re OCRing in reading. So you can never get to zero in, in terms of error rate. Mistakes will happen, but I believe computers do a better job than humans at this stuff, especially when humans get under pressure to go faster. You know you’re in a renewal season half of your team just got covid, whatever it is, right? When pressure happens, error rates go up too. Computers are always consistent. They’ll always make the same errors. Yeah. And they’ll always do the same things, right?

CW: <Laugh> Yeah. Failing the same way every time is, is, I don’t know, it’s comforting. You can build a process around it. 


Yeah, and another argument around it though are trying to shift away from the, the, the cost side of it. And, also considering some of the other costs, right? Like you can, it’s easy to look at and say, well, you’re paying somebody $8 an hour, that’s so cheap. But then stop and think about where that $ 8-an-hour worker is working. They’re working in an office space that’s in a footprint in a, and probably in a, in a, in a major city. And you know, what if you didn’t have all that overhead, right? Or you can greatly reduce that overhead. But then there’s also looking at it differently to say like, well, you know, you’ve had all these people you know, working in your processing center for the last you know, let’s say eight years or something like that.

Chances are they’ve learned a thing or two about your business along the way. Yep. What if you can use them in higher valued ways? So now the shift and the conversation goes from cost to from, from cash you know, cost to opportunity cost. It’s like what upside could we gain by having these, these, these people do higher valued work. And, and, and so that that opportunity cost side is something to look at. It’s a, it’s a harder of, it’s a harder sell and you usually have to be in a position where you just need the talent to grow your business. Right.

MG: Stacy you hit on a number of points that, that we’ve actually talked about over, over a few podcasts. Specifically that, that opportunity cost and, and helping automation, helping get people to do the, the work that it, I don’t wanna say it’s more meaningful, but that is more analytical, right? Like you’re, you’re taking out the, the data entry and the, the repeated steps to be, to be more what is the business outcome or how can they drive to specific results. I’m curious in, in your experience, what, what pieces of those different processes do you see that opportunity cost being most beneficial? Is it faster response time on a claim because you’ve got all that data there and so that person spends less time figuring out what the claim is and more time customer facing. Is it faster time from quote to bind because that data is getting in there and it’s cleaner on the underwriting side? Like where, where have you seen, I guess where have you focused more of the, of your time or where have you seen carriers that you’ve spoken to really focusing in within the different stages of these processes?

SB: I mean, you know, there’s definitely from, you know, wearing an operations hat quantifiable metrics that every, you know, service center, call center, back office team manages to, and you know, it, it, you know, volumetric ones like how many per day, how many per hour? There, there could be quality ones around around error rates, the number of times things go back through the process. So, so those, those measures are all, all there. I think one thing that doesn’t really get, you know, the analytical work is definitely something that you know, freeing up those resources to do more of that is, is definitely a piece of it. But there’s actually something very important that, I don’t know if, if you guys have been running into this and having, and seeing a lot of people talk about it, but the, the technology is, is actually enabling us to be more human again, right?

Less time spent having to enter data and even having to analyze that data by crunching it into reports and put into the spreadsheet and into that spreadsheet. Some people look at that as analytical stuff as, but, but what about being able to spend more time with your customers, show better empathy? What about actually having humans answer the damn phone, right? <Laugh>, like, geez, who does that anymore? Right? Yeah. I think, I think there’s a huge opportunity for, for for all this technology to enable us to just be going back to human again and spending more time together, building relationships, collaborating, showing empathy and, and, and and building trust with, with, with our customers, with our partners and, and, and others.

CW: Yeah. That, that’s a sub theme of this podcast, which is how human-centric, the whole automation world is and the, the opportunities that it unlocks for people. I want to, speaking of people, I wanna go back to another point you made, which is around you know, for some of these AI systems, especially the amount of data you need to train them to get to a high quality, and you mentioned the notion that there are vendors out there that are combining training data from multiple carriers. Why are carriers okay with that? I don’t, I don’t get it.

SB: Why would they not be? Okay, well, I can answer that one myself, right? <Laugh>. Okay, great. So because there’s a greater outcome for the whole in doing so, right? And it’s not like there’s a line around the block of, of carriers saying, I wanna sign up for that idea, right? But sometimes you know, companies have been smart around how they’ve put their licensing agreements together and how, and who owns what in terms of ip. And so the models are something that they will claim IP rights to they being the, the providers, right? And those, once those models get trained they own those models, and now the next company that they go to you know, they’re coming into the ta, they’re coming to the table with those models and they’re now improving them with new data sets. And, and, and, and that’s, that’s how the evolution of the models and the, and, and particularly the specialized models, right? That’s how, that’s how they improve. And one carrier can’t really do that themselves, cuz the only data they will ever see is the data in the four walls of their organization.

CW: Yep. No, that, that makes sense. Hmm. It’s interesting at some point, at some point, if a, if a tech vendor gets enough of this data, then, I mean, that’s a danger to the carriers themselves, it seems like. But, you

SB: Know, well, well sometimes, well u usually the few things, number one, if carriers are sharing data for, to the vendors, they’re, they’re, they’re usually ano they’re, they’re always anonymizing it. No vendors should ever wanna take data from a carrier that’s not anonymized, right? In one way or another. Whether it’s just outright in encrypting key fields, which makes it hard to learn or just using, you know, scrambler routines who, you know, will every, you know, every, every it’ll just randomly replace Chris with Stacy and whoever else is in the database and kinda switch ’em around. So you actually have real names in the fields. They’re just not, they’re not the right when you put ’em all together, the real people, right? So like there’s all kinds of routines around that, that for, for anonymizing. So but the, the, the, the, the, the thing is that usually the, the carriers aren’t like giving the data to the, to the provider, the providers bringing the solution to the data. Yeah. And then the models get trained and then and then they own it.

CW: Okay, good. Thank you for unpacking that for me. That’s helpful.

SB: Sure. 

MG: I wanna pull on this thread of data just a little bit cuz something else that, that Chris and I have talked about is one, the data that the carrier or, or the broker or the reinsurer has in-house, as you mentioned, Stacy, right? And then there’s a lot of of companies out there now selling data you know, as a service, right? Mm-Hmm. <affirmative> and, and we’ve seen a lot of companies doing this to, to enhance or supplement the submission process or you know, validation that, that what’s on the submission or even what’s on the claim is is correct. Ha have you seen a a lot of, a lot of solutions or a lot of focus on, on additional data to support or supplement existing carrier data? And what are some examples there?

SB: Yeah, that example let me see if anything comes to mind. Sometimes my brain freezes on these specific examples, but definitely, I’m I’ve seen, I’ve seen lots of solutions and I, and I know all the carriers are working on it. Like, for example InsureTech Hartford has an annual innovation challenge where we work with carriers to identify themes and challenges that they have, and then we’ll, we’ll do an open call for solutions. And last year the 2022 Innovation Challenge, the Hartford sponsored a category around data for, IM improving the underwriting process, right? So that’s just one example that comes to mind of carriers being hungry for this. And there is a ton of data out there, and like at at my last company, we had a person whose job it was just to constantly be looking at the marketplace for all the data providers to figure out who’s bringing in something that’s new and different, and every line of business kind of has its own needs.

 And so, you know, there’s, there’s a, there’s a few data providers out there that are like the giants of the world, like let’s say the, the virus and the Carpe Datas and the LexusNexus and folks like that. And they’re all gonna love me for saying that. And anyone I forget is now hating me, but <laugh> <laugh>, but, you know, there’s definitely the big conglomerates, but there’s very niche providers out there as well. And and I worked for a global carrier, and one of the, the, the challenges there was the data set consistently across the globe wasn’t there, right. You’d be able to get, you know, much level, much better granular data about, you know, property in the United States than let’s say, you know, property in Uganda, right? <Laugh>, you know, and, and so your processes have to be tuned to, to that as well, right? So we like to think about, boy, if you just do everything the same way, it doesn’t always work that way.

MG: And, and I’ll, I’ll just add onto that, and Chris, this is one of your favorite things to, to just like digest, is sometimes the data you want is just not available due to the regulatory environment in that country, right? So I, I, the one example I know of is the amount of data available on small businesses between what’s available in the UK and what’s available in the US is strikingly different just based on what these businesses have to do from a self recording perspective. So, Stacy, to, to your point, I’m sure there’s, even if you have a strategy, right, a data strategy or it to implement that globally must look different based on, on your area, your geography of focus.

SB: Yeah, absolutely it does.

MG: And then from a technology standpoint, it must pose other challenges from a how do you get the system ready? How, how do you make sure that you have different systems that can take in whatever’s available or pass through on something if, if something’s not available, that that exists in a different process?

SB: Yeah. Nowadays, you know, architectures are very integration, you know, focused everyone. Every, every carrier has to be good at doing integrations Now to not be able to do integrations would be a problem, I’m sure. And, you know, there are, there are a lot of different patterns for, for how to do that. You know, we’ve seen a lot of event-based architecture type of solutions becoming, you know, more and more popular which allows things to basically move at the speed of, of, of, of computers, right? <Laugh> as opposed to, you know, in, in the old days everything be batch scheduled and you know, the, even even when APIs came along, people would batch schedule their APIs, right? And so, so now, you know, between, with cloud computing and allowing people to scale up and scale down more, more easily as, as well as these, these more modern event-based architectures like a Kafka or I forget what Microsoft’s version of that is, but the, you know, these types of, so solutions are enabling us to do more integrations faster and still be performant.

 Cuz in, in the olden days, it’s like every time you added another step into the thing, it was like another, you know, three seconds that someone was gonna be waiting. And so there was always that challenge. So we, so in addition to all the data we are seeing, the, the technology advances both in compute, power storage and architectures and security, all these things are just amazing the the way that solutions are being built then let’s say, you know, 20 years ago when I fir when I used to do it myself by hand, right? So <laugh>

CW: Yeah, it’s also changing the cost structure as well. But we don’t have to, we don’t have to go into that quagmire just yet. You, we’ve talked a lot about the challenges and what’s hard to automate and why it’s hard to automate, but then you, I was getting bummed out honestly, but then you raised, you know, the fact that integrations are making a lot of things simpler. So on the automation side, especially when it comes to, you know, artificial intelligence, that kind of framework, what’s working actually on these, on these intake processes?

SB: Hmm. There’s, it seems that you can get to like 80% pretty quickly, right? Yeah. Okay. 

CW: 80% of what

SB: Actually, I, I was just gonna say by 80%. I I would say in two dimensions, number one, 80% of the data elements that you’re looking for can get you 80% accuracy. Yeah. you know, pretty, pretty quickly. Then you wind up having these certain data elements that become difficult and harder to find. I’ve actually seen one of the things that’s always been a challenge is extracting financials from financial reports. Yeah. Isn’t every, everybody has a different way of doing it. So the, the old days the solution was, well, you know, we’ll just go to like capital IQ or, or, or you know, one of those, you know, s and p or one of those data providers because they had already conformed everybody’s financial reports into some common data structure, right? But, but now what we’re looking at are solutions that yeah, you could take the annual report and and suck it in, and we, we know how the assets, liabilities, and equities break down and and we can get all the ratios that, you know, that we want out of it once we have that.

So that’s, that’s advanced r that’s, that’s stuff that I’ve seen people really struggle with and I’m excited to see please don’t reremember ask me to remember the name of the company, but I, I recently seen somebody that can actually do that pretty well. And that is a part of the intake process, depending on the line of business, you’re, you’re writing anything, a lot of the, and a lot of the things that are in, in commercial and liability related are gonna require financial statements, right? And and, and so being able to extract data from financial statements is a, is a huge challenge. But anyway, 80% of 80% pretty quickly. One of but then in terms of accuracy, you, you, there’s that point of diminishing returns that remember you’ll never hit a hundred percent, even if you are a hundred percent human, you will not be a hundred percent.

So there’s with anything quality wise, it it is, it’s, you reach these points where it’s a lot more investment for not as, not as big of a, of, of an improvement. And if anybody’s getting above 95% accuracy you know, they’re doing pretty damn well. And and so that’s another thing. One of the other challenges though, when you start taking re you know, truly unstructured stuff like out of the mail and or, or even looking at, at certain co contracts you know people start like highlighting things and color coding things and drawing arrows on the form. You know, like, I meant this, well, when this happens, circle arrow over here, insert over here, right? And it’s like, how is a computer supposed to read that? Like, there will always be a, like, I think we’re still generations of technology away, which by the way is faster than human generations, but we are still many generations away from computers being able to deal with some of those things. But when it comes to reading letters reading structured forms, there’s a lot of solutions out there for that stuff right now. So the question really that I would have for you if I can flip it around is what, what, what differentiates companies now, right? Because if everyone’s using basically the same, you know, frameworks and same patterns you know, what’s, what’s really differentiating companies?

CW: I’m gonna, I’m gonna let Michelle a answer that in the more general case, and then I, I can talk a little bit about Indico specifically.

MG: Yeah, I think you mentioned it Stacy, at, at, at one point of eventually something will break down if, if it’s rules based or templates based. I think what, what what we’re seeing is again to that, that ability to, to specialize what are the solutions that are leveraging AI or machine learning to not just have something fall into, to a rules-based engine, but something that is more adaptable and flexible to, to how things come in. Which is probably a nice tee to, to Chris, but I’ll make one other point as well, is that before it used to be we have AI or we have a solution that can do anything, just tell us what you wanna do and we can figure out how to do it. And I think what, what we’re seeing now, to your point earlier is I can come in and do X very well, and I can automate that process, or I can solve for this, or I can make sure that all of your data is valid here.

 And so the, the specialized solutions that, that maybe is get, get in the door prove out the, the accuracy and the validity, and then build from there and continue the automation either on the claims or the underwriting side. So I think what we’re seeing doing really well is the proven business outcome on something very specific that’s solving a true pain point, and then expanding from there as opposed to being, I have a technology solution that will solve all your problems. Just tell me what they are, that that doesn’t seem to resonate too much. <Laugh>.

CW: Yeah, I a hundred percent agree with all of that. I I say this a lot, I’ve never said it out loud on the podcast and who knows how marketing will feel, but I don’t think Indico actually has a tech advantage. And I think anyone in this space, you know, in the language modeling sort of the unstructured data space that says we have a tech advantage, and they’re not like open AI <laugh> or, you know, Microsoft or Google they’re probably selling snake oil. The thing that differentiates solutions to me nowadays is have they actually solved the problem? Have they learned from solving the problem in the enterprise? And then has all of that learning gone back into the product? And so, you know, Indico got to the language modeling space way before anyone else, and it was, it was too early. But one of the advantages to that was we, we stepped on every landmine there is before anyone else did.

 And so the product is really built for the business user to be able to come in, essentially do their job and have the machine learn from them doing their job. And then, you know, the parts that work, it’s very clear, which, what the parts are that work and the parts that don’t. There are ways for those to, you know, flow through the process and get to the right people and, and the right subsequent downstream remediation efforts. So yeah, that, that’s, I don’t, I don’t think anyone has a, a tech differentiate differentiator except in the sense of like user experience and understanding of what people are trying to do. And what’s hard about it,

SB: You know, I think you just differentiated yourself, Chris, by being open and honest about that. Oh, <laugh>. Cause I, I, I, I, I, you know, I somewhat agree. Like, I’ve tried to be kind about it by saying there are only so many frameworks out there for this type of stuff. And from a tech perspective, everyone’s making the same thing, the implementation patterns. There are a few ways to do that. And everyone’s pretty much doing the same thing. It’s, it’s not what you have, it’s how you use it that’s making the difference now. Right?

CW: Yep. Absolutely. I couldn’t agree more. Well,

MG: I will end up complimenting Chris on that one. So Stacy, Stacy, thank you so much for taking the time to chat with us today. It was insightful and enlightening. We appreciate you joining us. So thank you. And for everyone else, this has been another episode of Unstructured Unlocked. Thanks for joining.

SB: Thanks.

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

Subscribe to our LinkedIn newsletter.

Unstructured Unlocked podcast

April 10, 2024 | E44

Unstructured Unlocked episode 44 with Tom Wilde, Indico Data CEO, and Robin Merttens, Executive Chairman of InsTech

podcast episode artwork
March 27, 2024 | E43

Unstructured Unlocked episode 43 with Sunil Rao, Chief Executive Officer at Tribble

podcast episode artwork
March 13, 2024 | E42

Unstructured Unlocked episode 42 with Arthur Borden, VP of Digital Business Systems & Architecture for Everest and Alex Taylor, Global Head of Emerging Technology for QBE Ventures

podcast episode artwork

Get started with Indico

1-1 Demo



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

Unstructured Unlocked

Enterprise leaders discuss how to unlock value from unstructured data.

YouTube Channel

Check out our YouTube channel to see clips from our podcast and more.