Emtech Group Podcast
Episode #6: XML Validation for Accurate Data Integrity in Insurance
Summary
In this podcast, the speakers discuss the challenges of maintaining quality assurance (QA) within insurance carriers. One key point is that while many companies strive for high quality, few make it a core competency. QA is often left until the end of projects, especially in smaller teams with limited resources, leading to more manual testing and the potential for errors. Even companies with more developed QA processes may struggle with automation, which can affect the accuracy and reliability of their outputs. The conversation emphasizes that quality should be seen as a strategic investment, not merely a required task, to enhance operational efficiency and avoid critical issues.
The discussion also focuses on the importance of data integrity in the insurance sector, highlighting a new product, QMT TrueXML, designed to validate and test data. The software ensures files are accurate and complete, allowing companies to catch errors earlier in the process, thereby saving time and reducing operational costs. Automation in QA can help minimize errors and prevent business disruptions, protecting both internal processes and external brand reputation. Proactive QA, they conclude, is essential to prevent inefficiencies and improve customer experience, while also maintaining strong business performance.
To find more episodes of Emtech Group Podcasts, visit our resource center. To read more about Quality Assurance, QMT, QMT TrueXML, and technology topics, visit our blog.
To find the transcription of this podcast, scroll to the bottom of the page.
The views and opinions expressed in the podcast are those of the guests and do not necessarily reflect the official policy or position of Emtech Group Inc.
Featuring
Kevin Kelly
Kevin is a former Fortune 100 executive and currently serves as a strategic advisor for CIOs and their leadership teams. He is recognized for his expertise in organizational and technology transformation, product development, and aligning technology strategies with business goals. His career includes leadership roles at major organizations where he successfully delivered large-scale transformation programs and optimized technology initiatives. Kevin holds a BS in Computer Science and an MBA from the University of Texas.
Gary Tyreman
Gary is passionate about software products and their impact on businesses. He has over 25 years of experience in the software industry, including guiding a company turn-around and building a top-3 global high-performance computing software company. He is a seasoned leader in change management.
Transcript
00:00:18 Gary Tyreman
Hey, Kevin. It’s good to see you again.
00:00:20 Kevin Kelly
My pleasure.
00:00:21 Gary Tyreman
Think last time we talked; it was the dog days of summer. Now here we are, right in the middle of. Well, I guess the beginning of fall, so to speak.
00:00:28 Gary Tyreman
Everything’s well.
00:00:29 Kevin Kelly
Yeah, everything’s going well. We’re looking forward to some cooler weather here.
00:00:32 Gary Tyreman
Yes, absolutely. Well, the Farmers Almanac has predicted my area is going to be colder and more snow than usual, but Canada, what can you do? Happy to share that.
00:00:43 Kevin Kelly
Yeah. Well, I’m in. I’m in Texas, Gary. So, we will take cooler and snow any day. We can get it.
00:00:49 Gary Tyreman
Yeah, well, you know, Austin shuts down when there’s a little bit of snow. I’ve seen the fly over, shut down and it’s like, really come on.
00:00:55 Kevin Kelly
Yep.
00:00:57
It.
00:00:58 Kevin Kelly
It doesn’t even require actual snow; they can just forecast snow, and we close the schools and the roads.
00:01:03 Gary Tyreman
Yeah, yeah, I’ve seen that. I’ve seen that.
00:01:06 Gary Tyreman
So let me want.
00:01:07 Gary Tyreman
To pick up where we left off conversation, we had a little while ago. I was drilling into your experiences as a CIO from New York Life, and if you had any advice around, you know, for smaller organizations, IT executives or managers and.
00:01:27 Gary Tyreman
You did. You had some great words about.
00:01:30 Gary Tyreman
Take.
00:01:30 Gary Tyreman
Seeing quality as a discipline and creating a competence around this right, we’ve talked before about QA’s not a core competence. Generally speaking, I don’t want to exclude those who have been successful at making it, but I thought that was a, a very good.
00:01:46 Gary Tyreman
You know, piece of advice, we also talked a little bit about the processes and the data flow.
00:01:53 Gary Tyreman
In between them.
00:01:54 Gary Tyreman
And I think you have seen the news. We just released our product which does data validation and integrated into our QMT our model-based approach for insurance carriers. So, I kind of wanted to drill a little bit about that and get your get your thoughts on that.
00:02:12 Kevin Kelly
Yeah. Congratulations. That was a great announcement to see coming out. And I know you guys have been working hard and long on that product release and congratulations, we’ve been, we’ve been watching you in the industry and I think it’s a great development and will benefit your clients.
00:02:32 Kevin Kelly
Very well. So. So I do think, Gary, you know, quality is not necessarily competence or competency that a lot of carriers have. I think everybody wants to do a good job at quality, but it’s not something that is consistently focused on you have carriers.
00:02:33 Gary Tyreman
And thank you.
00:02:53 Kevin Kelly
All into the spectrum you have, you know, smaller carriers that may have small teams.
00:02:59 Kevin Kelly
And QA is something that that gets done at the end of a project and you may have carriers’ kind of in the middle of the spectrum that maybe they have a robust QA life cycle or robust methodology around it, but they may not have a lot of automation. You know maybe it’s a whole lot of manual testing and you know maybe they test the right things or maybe something slips through.
00:03:21 Kevin Kelly
Because the test suites aren’t very mature in a handful of cases, and it might be 5 to 10% of carrier.
00:03:29 Kevin Kelly
There’s have a very robust quality process and protocols in place, but regardless I think of where an insurance company is on that spectrum, QA should, and quality should be considered an investment in something that you invest in versus something you have to do or necessary evil. And whether you’re small.
00:03:50 Kevin Kelly
Or big whether you’re very mature or not as mature. If you come at it with the right mindset, I think you can really turn the whole quality.
00:04:01 Kevin Kelly
Thing and to not something you have to do but into something that becomes a strategic enabler for the technology organization and therefore for the business overall.
00:04:12 Gary Tyreman
It’s a very, very good point, very good point. So, one of the questions I had for you is in your experience with QA, either through your clients or your own personal experiences at different carriers or working with different ensure techs, the challenges that we seem to.
00:04:30 Gary Tyreman
Find which is why we spent time figuring out how to design this product as an extension to QT. QMT True XML is the name of the product and what it what it does is it integrates into the model the configuration that a user would provide of their XML file and then compares.
00:04:51 Gary Tyreman
The expected results versus the actual results in the file for each successful test case, so if you’re.
00:05:00 Gary Tyreman
Randomly generating data and you’re putting it through the model. Each test case with different data would be validated is it is the file complete and is the file accurate. Complete meaning is everything that’s supposed to be there and accurate from the point of view just because it’s there doesn’t mean it’s correct. So that’s.
00:05:20 Gary Tyreman
That’s the capability at a high level of the product and what we’ve seen is a lot of people and this is my question, if you agree or if you have thoughts on it.
00:05:29 Gary Tyreman
Don’t they seem to skip testing data integrity?
00:05:33 Kevin Kelly
Yeah. And I don’t think that’s intentional, Gary. I think it’s unfortunately a byproduct of the fact that a lot of this testing is done manually still. So, you’re not able to really do the boundary testing, the edge testing all the permutations.
00:05:54 Kevin Kelly
What’s valid? What’s not valid, and you know, because that’s all done manually you. It’s not that necessarily it’s skip.
00:06:02 Kevin Kelly
But there’s a lot of blanks in it. There’s a lot of blind spots and you know, in some cases, you know, holes in it, you know, that are relatively large, and you know, to the degree that companies can automate that stuff. And that’s one of the beauties of QMT and the product that M tech is offering, is it minimize that.
00:06:23 Kevin Kelly
And may not 100% eliminate it, but to a very significant and strategic degree that addresses those kinds of challenges.
00:06:33 Gary Tyreman
So, what would? What would you call, right, so when you’re when the when that file is formed its transient right between systems.
00:06:40 Gary Tyreman
And when it’s formed and by formed, I mean it’s being written, written out the data in it, if it’s incomplete or inaccurate, what do we call that? Is that like deficiency or is that deficient data, invalid data? Like what? What is the word you would use to describe that?
00:06:56 Kevin Kelly
So, you you’re talking Gary, like if you if you have an EF system that maybe has partial data in it but hasn’t.
00:07:03 Kevin Kelly
Hasn’t completed it, no.
00:07:04 Gary Tyreman
Yeah, for example it, you know, it may have an XML tag that’s incorrect, or it could be missing. It could have the tag, but it’s missing the data in the tag. So, what’s being what’s expected there isn’t. Isn’t actually written, but the file is assumed to be complete so that that could be enough. That’s a writing error or formation error.
00:07:24 Kevin Kelly
Yeah, yeah, yeah. And in that case, you know, at a minimum, you’ll catch it on the back end when.
00:07:29 Kevin Kelly
That’s submitted, right? But you don’t want to catch it there. You want to catch it? As far as close to the front end of the process as you can. So that, that’s where automating all of those different test criteria and you know what? What, what, what does success look like? What’s the definition of successful outcome and getting that automated is?
00:07:49 Kevin Kelly
As close to the origination of the data as possible. That that’s important. If you can catch it at the at the time you’re doing the data entry, you know or at the time someone has created an application.
00:08:03 Kevin Kelly
That is a very easy and cheap way to find and fix the data problem right there. If you wait until it gets submitted down the line and it falls out of some QA process, you know hours or days later and then you have to circle back, you’re wasting time. Your cycle times are increasing your operational.
00:08:23 Kevin Kelly
Costs are increasing. You are more likely than not have to get human beings involved to.
00:08:29 Kevin Kelly
Deconflict or fix the issue so the longer it takes you to detect the data problem, the more expensive it is to address it and to the degree that you can automate that stuff on the front end. Not only is it a better you know from an operational standpoint it’s cheaper to do, but if the customer experience.
00:08:51 Kevin Kelly
There, the user experience is a lot better.
00:08:53 Gary Tyreman
So, if it’s missed in the in the early points of processing, right the creation and it slips through my words, certainly not a technical term, but if it.
00:09:04 Gary Tyreman
Ends up getting pushed into policy administration and that policy admin system runs on batch, so it may not actually be available for a day, possibly two. It depends on their curiosity. They would find the air at that point by doing a light bulb test with the PDF printed and.
00:09:14
Hmm.
00:09:24 Gary Tyreman
Is that kind of what they would do and that is that?
00:09:26 Gary Tyreman
How would they discover it?
00:09:27 Kevin Kelly
Yes, sometimes it.
00:09:29 Kevin Kelly
We discovered that way a lot of times it will also fall out of the business process, the back-end process. When you go to submit an application or move it, move it further down the process. So maybe move it to the next stage of underwriting or so forth, it’ll fall out of a workload also. So, you know so you can have.
00:09:48 Gary Tyreman
OK.
00:09:50 Kevin Kelly
Some workflow errors or process errors down.
00:09:53 Kevin Kelly
Stream as well, and that that’s probably going to happen before you get to that kind of light bulb test. You know where your kind of holding the policy up to the light and saying do I have all the right data; it’ll probably fall out before that because something in the process is going to break because it doesn’t have all the right inputs to it.
00:10:12 Gary Tyreman
That that makes sense, right? So, it would, it would fall out. Is there an analogy to the light bulb test, right? So, if I’m looking at my documentation and I hold things up, is there an analogy in the in the in the data, because we’re talking about XML files here, right, possibly, right. And the verification validation?
00:10:29 Gary Tyreman
If I’m practicing some QA and I’m doing manual observation, is there an analogy to that light bulb?
00:10:36 Kevin Kelly
Thought process. Yeah, I think I think there is Gary, right? So, but you know back in the day you know people would hold up you know, you know I’ve gotten a piece of paper. I want to hold it up to the light and I got the model. I know it should look like this. Does it look like.
00:10:49 Kevin Kelly
That you know, when there was that kind of kind of smoke testing, so to speak, if you will, like does.
00:10:55 Kevin Kelly
Look the way it should look, and I think today with tools like QMT in the market, whether it’s an XML file or even a PDF or different types of data files, you have a good model. You know what it should look like, you know what data elements are required, you know which ones are optional.
00:11:15 Kevin Kelly
You know what data configurations are allowed and what you’re not, and these tools like UMP can automate that in in a very robust, repeatable fashion that will get better outcomes for the carriers.
00:11:29 Kevin Kelly
All across the process. Excellent.
00:11:31 Gary Tyreman
If I have malformed data pushing through my system, it hiccups in a back-end workflow. There’s an error. There’s a there’s a business cost to that, right? Not just the technical issues, but what it what is? What is the business cost of that? Like, is that an air that embarrasses an organization or?
00:11:50 Gary Tyreman
How would that be communicated internally?
00:11:52 Kevin Kelly
If you’re, if you’re lucky, it is an error that can be addressed internally with no.
00:12:00 Kevin Kelly
No adverse consequence other than the fact that you have to spend cycle times, and you have to touch something more than once. So, it impacts your operational efficiency in a negative fashion for sure. But hopefully that’s all. If you’re lucky, if you’re not lucky, depending on the error it becomes visible.
00:12:20 Kevin Kelly
To.
00:12:21 Kevin Kelly
Your distribution, your agent or the broker selling the business that that can have negative brand damage. You know you obviously want to be easy to work with for companies, insurance carriers that have captive agencies or captive agents, OK, might be a irritation but may not be a big deal.
00:12:42 Kevin Kelly
But if you’re if you don’t have captive agents and your agents can choose the cell with one carrier or another if you’re too hard to do business with and you and you keep breaking, they’re gonna quit doing business with you or quit doing as much business with you and they’re going to go to the other. And so, you start to lose.
00:13:00 Kevin Kelly
Business and yet another scenario there is if it becomes a problem that’s visible to the policyholder, that’s maybe even worse, right, because now you’re impacting the actual end consumer and the and the purchaser of the product. So, there can be a lot of negative consequences.
00:13:20 Kevin Kelly
Getting QA wrong, getting the data wrong audit and the absolute best outcome when it’s wrong is it’s purely internal and you.
00:13:29 Kevin Kelly
Just have a process and efficiency or process problem, but a lot of times a really bad day. I’ve been involved in product rollouts before where there were scenarios that weren’t tested correctly, and you know how we found out about it. We found out about it because the agents called the call centre, and you know that’s how we found out so.
00:13:49 Gary Tyreman
Yes.
00:13:49 Kevin Kelly
That’s a bad day when that happens.
00:13:51 Gary Tyreman
That’s.
00:13:52 Gary Tyreman
The that’s the unhappy customer in this case, so the. So, if I if it affects operational efficiency clearly that causes delays cause you got to do it again. But would that also be considered like a hit on margin because something if it’s a live system, I imagine it would be whereas visible to distribution is brand impact.
00:14:12 Gary Tyreman
Right. And that can also be lost, lost revenue, lost business. So is it, is it a matter of one side internally if it’s discovered internally, you’re touching it more than once, it’s a margin to it’s a margin discussion. But if I’m hitting it with my distributors, I could have reputational damage.
00:14:28 Gary Tyreman
Terrible consequences around. They’re not coming back. Lost business, right? Is. Is that kind of how it would be measured? So, the CEO is going to be getting updates on. OH, we lost another distributor.
00:14:39 Kevin Kelly
Yeah, that’s right. It does have an impact on margin, but you know typically the CEOs of the world, they’re not, they’re not measuring our margin but they, but they do measure on operation metrics. So, it will show up.
00:14:52 Kevin Kelly
In in the form of adverse operational metrics there ultimately that does impact how many employees do you need to stamp X number of widgets, right? And so, it does translate into increased variable cost or maybe even increased fixed costs which would impact margin.
00:15:12 Gary Tyreman
I got it. Got it. Got it. So, I think the moral of the story here is test, test, test again. Be proactive, right. You talked about the discipline mindset and looking at it as an investment and a strategic.
00:15:12 Kevin Kelly
On that side, yeah.
00:15:29 Gary Tyreman
Enabler. If that approach is taken, many of these problems that we were just talking about can be avoided, minimized as maybe the more accurate way of saying that all right.
00:15:39 Kevin Kelly
Yeah, yeah. And you’ll never be able to truly eliminate them. There will always be something that happens, you know, something always goes bump in the night. And if this stuff were easy, everybody would be doing it. But you know, the good operational or technology executive is going to try to be as proactive as they can.
00:15:59 Kevin Kelly
To work, shift quality as far to the left as possible, automate as much as possible and it’s not just about doing the initial automation, but it’s about rinse and repeat every time you do a product release or if you do a minor update to a system, you can rerun the cycle and rerun the testing.
00:16:19 Kevin Kelly
And.
00:16:20 Kevin Kelly
You can sleep well at night knowing that you did everything you can to prevent any adverse outcomes.
00:16:26 Gary Tyreman
Understood. Well, I think we’re at the top of the time. So, I want to thank you again, Kevin, for a great discussion. It’s always insightful. It’s always good. I love learning more and its always fun right to talk to the people who’ve actually done right, been in this seat, so to speak, right? So, I appreciate it. We’ll have to do this again.
00:16:45 Kevin Kelly
Yep, absolutely, Gary. Thank you and enjoy the enjoy the cooler weather in Canada.
00:16:50 Gary Tyreman
I appreciate it. Thank you.