DAVID LEDGERWOOD: It’s great to have you. Thanks for joining.
DAVE SHEWFELT: Thanks, Ledge. It’s great to be on.
LEDGE: If you don’t mind, maybe give a little background story of you and your work, let the audience get to know you a little bit.
DAVE: Sure. I have quite a bit of a varied background for a lot of tech folks. My initial after-college work was 12 years in the Marine Corps. That’s definitely an interesting leadership and learning experience. I definitely don’t regret it at all.
At some point it came time to get out. I started doing e-commerce. Technical . CTO of a small e-commerce startup for a while. Then decided I needed to make some changes, so I moved back to the East Coast and that’s where I got involved in my current role.
I started as Solution Architect. I’m now the Director of Integration Services for Vidado.ai – or if anybody has heard of Captricity, we recently rebranded.
What we do is, if you want to put it in the modern buzz phrase, is the intelligent automation. But really the core of that is this idea of capturing handwriting off of forms at a level that’s equal to or better than a human being, using artificial intelligence and machine learning. Quite fascinating stuff.
We deal a lot with highly regulated industries; insurance, financial, government institutions. I don’t know what the exact number is but I think that many of us, especially those of us in the tech world, would be shocked at how many people in the United States every day do not get on the internet, don’t use a computer. There is a lot of the highly regulated industries that still require paper. These places have a lot of forms that go around.
In addition to that, a lot of these places literally have just like aircraft hangers full of paper, floor to ceiling, of data that’s never been captured. It’s literally a gold mine of paper.
We call those more like the analytics type stuff, and we get some of those. Then we do a lot of the operational stuff. Literally, as people are sending their forms in today to change their beneficiaries, we’re processing that and extracting messy handwriting off of very poor quality scans from faxes and things and like that, and returning digitized data to them to be acted upon operationally.
So, a lot of really cool stuff.
Again, I’m on the integration side and we can definitely get into some of the interesting stories there, but I’m not going to take credit for the machine learning team. We’ve got some just unbelievably brilliant people getting patents and stuff like that on this ability to read human handwriting off a paper.
That’s where I’m at right now.
LEDGE: Wow. I looked at your resume and I was thinking about, you worked in intelligence and there’s got to be all kinds of interesting signal SIGINT types of stuff there. Did that play at all into wanting to go into such a data rich field, or is that just like a happy accident?
DAVE: Well, from the time I was a kid – and the first time I ever had my dad take me to a recruiter was the 7th grade. I always had this mentality of military, but I love to build things. My initial thing was, I wanted to be a combat engineer because their logo was a guy with a shovel and a rifle. It’s like, I can build something and I can shoot something.
LEDGE: Shoot stuff.
DAVE: Yeah. But after college – in college I was electrical engineering before computers were really a thing, so I’m kind of aging myself here – I got in the Marine Corps and the signals intelligence field was, again, that combination I love.
Look, I’m not going to lie. I can’t at this point. I’m a complete geek. But I’m in the Marine Corps and I love that too. So, it’s just like that’s the perfect match for me.
Certainly something like the SIGINT field in the Marine Corps is going to be one of the more technical fields for something that’s not typically considered a technical organization. But a pretty technical field. We got to use, at that time, a lot of interesting equipment and stuff, and actually get into some of the computer side of it as time went by.
It was a good combination, and it just fueled. If you talk about pivotal moments in your life, I can remember on my thesis project for my graduate degree, having to put together a network of 50 computers. Now, I’d never really done that. I worked 36 hours straight locked in a building because I was fascinated by it.
From that point on – it was back in the early days of Red Hat where you had to send away and get it on the disk and send it to you and stuff because you didn’t have fast enough speed to download it all. All that stuff. It just grew from there.
When I got out, I hooked up with an e-commerce startup and I've never looked back. It’s such a wide range. That leadership side of 100 people in an organization, leading that, being responsible for it. Down to staying up late at night coding and learning a new programming language or something like that. I kind of get the best of both worlds.
LEDGE: Scratch all the itches. That’s a pretty good deal. Not everybody gets to say that.
I think there’s a fascinating body of literature and work around the armed services disciplines being brought into civilian work, and the leadership lessons. All those things that most of us have not had access to.
How do you draw on that on a daily basis? You go back and you’re like, that’s your regular career in life and that’s what you did, but there’s so much to learn there and impart upon an organization.
I’m curious. Key lessons there. What can the rest of us take away, having not had that conversation?
DAVE: That’s a great question. I’ll tell you one of the fascinating things. The first time I ever had somebody start to go through the whole agile/scrum process, all the rest of it, I just kind of looked at it and I was like, we never called it that but that’s essentially what we did in the military in any of our stuff. You figure out where you’re at, you figure out where you want to go, you figure out what step you should take, you take it, you reassess.
Literally in the military, I think John Boyd was the guy’s name from the Air Force, they called it the OODA loop – Observe, Orient, Decide, Act. You just keep repeating that cycle.
He was an Air Force combat pilot and they came up with this OODA loop thing. I think it’s been picked up a lot in the business world, that you even hear that phrase. The point was, you win, you beat the other guy by doing your loop faster. A lot of that really boils down to this very agile mentality of, where am I at, where am I trying to go? You take a little step. You don’t jump all the way from A to B. You don’t win by being a zero tolerance. There’s going to be mistakes.
Beyond that, a lot of it is just good leadership skills in terms of people. What was it, people soft or something, they talk about almost no technical projects fail because of the lack of technology, they fail because of people issues. So bringing that into perspective, a big part is, don’t ignore the people part of it and take those little steps and reassess.
LEDGE: Absolutely. I remember doing after action reports, or what have you, and everybody has their own sort of lexicon for it. Coming out of a discipline of myself, coming out of continual learning and continual improvement and things like that, it always made a lot of sense to me.
Yet, and I don’t know if you’d had this experience, and maybe not with the software people, but I often find that for whatever reason agile is not the disposition that most folks come to the table with. I always wonder how it seems so logical, and it makes so much sense, and yet virtually everybody needs to reprogram their brain to get there in the first place.
I don’t know if you’ve ever given any thought to that, but it always happens that way. Call it waterfall. Why does that disposition come naturally to us and we have to have it kind of whipped out of us to start acting in the way that ends up being the most logical approach?
DAVE: Yeah. We deal with a lot of enterprise, Fortune 500 and above, companies that have been around 100+ years. A lot of that is just ingrained. Sometimes there’s a comfort level for people to just do things very strict and have everything defined, and then do the loop. It’s somewhat uncomfortable for a lot of people. True agile, that mindset of constantly reassessing, constantly changing, that’s very disturbing, I think, for a lot of people in normal life.
LEDGE: Maybe it’s that technology or business processes and such have reached a level of complexity where everybody, military or whatever, needed to realize that, we can’t figure this out because there's too many variables.
Maybe the complexity of business and complexity of process and things that we’re able to do with technology, reached such a critical mass that you could try to plan, you’re going to be so wildly wrong that it’s not worth it. So, let’s all acknowledge that reality.
I don’t know. It always struck me that it’s not the regular disposition, because you watch a little kid trying to walk, they’re not planning it. I wonder, when did that get beat out of us. How does that become not the way to go?
Tell us some integration stories. You must deal with some crazy stuff. I want to know how a whole aircraft hanger full of papers gets scanned in. I mean, forget about the fact you need to process what’s on the paper, scanning that much paper. What’s that look like?
DAVE: Thankfully, there are companies out there that actually do that. Thankfully, we don’t have to do the scanning because, yes, that would be a monumental task.
I’ll tell you. Actually, one of the podcast I listened, the guy, Steff Kelsey from Notarize used that word, transformational. So many things that he said struck such a chord because, again, that’s essentially what we’re doing. We’re very transformational. We’re going into companies that have been around 150 years and they have these processes, they’re all manual, and we’re trying to transform them.
You hear a lot today about digital transformation advertisements during football games about digital transformation.
I’ll tell you, I love this story because it was one of the biggest light bulbs that went off for me in the past couple of years of this. Is, a lot of times, not surprisingly, we get pulled into these companies that have these paper workbooks. They want to automate them. They need to be able to extract the handwriting.
Now, not surprisingly, kind of like the whole thing out of the movie script, we go in there and there’s always a little bit of this tension. You’re looking people in the eye who they essentially think that you’re there to remove them. You’re going to cost them their job. So then you try and start building this relationship, and over time – not a lot of friends and stuff just on a personal level just very friendly relationships but there’s always that tension at the beginning.
We start working with them, and one of the things – it took me a little while to pick up on this – but when our software, our SaaS solution, when our software would fail there wasn’t just this thing of turn in a defect report, there was this literal emotional glee. A personal emotional glee on the part of the individual doing it.
All of a sudden, the light bulb went off. Our failure is literally a vindication of their self worth as a human being in terms of doing the job. It was this complete bombshell of, oh my goodness, how critical the human element is in it.
Let me expand story, if you don’t mind. Here’s the next piece of it. The people might be thinking, okay, great, you’ve come up with this machine learning solution you can digitize the handwriting off a page. Okay. Good. Done. We just go out there. We just have everybody toss their papers in the hopper and off we go, we automate it. The problem is, no form, piece of paper with handwriting on it is really just that information. It’s part of a business process, business logic.
Ledge, I had an enterprise customer – if I said the name, everybody here would have heard of them – they literally said to me, “We don’t want what’s written on the piece of paper, we want what they should have written.” It’s not enough to just figure out how to get handwriting off a piece of paper, now we have to trace back through ESP and get what the person should have written on the piece of paper.
LEDGE: I've had this conversation with AI companies and professionals over and over and over again. What we’re doing here is creating technology that gives superpowers to those people in those seats. This is an augmentation and not a replacement.
LEDGE: Time and time and time again, maybe that conversation with whoever you’re talking to is worthwhile – and they probably don’t believe you right away, they probably have been fed that line before – but that is so true. They know so much tacit knowledge, there’s no way you could ever program that, ever.
DAVE: Right. It’s got to be a mix. I like the word you used, augmentation.
As a side note, so many times I've seen where these companies come in and they have this idea, kind of the old school mentality of, “I’ll just write a big enough check and then one day we’re going to have the ribbon cutting ceremony, and we’re going push the big red magic button, and we can dismiss all the people, and we’ll just be…”
When they take that all or nothing approach, usually what ends up happening is a lot of frustration, a lot of wheel grinding and spinning, and nothing at the end.
LEDGE: You’ll end up with a total disaster. You talk about the glee, all the people will be like, “Told ya!”
DAVE: It just entrenches them deeper into like, “Oh yeah, we tried that before and it didn’t work.”
LEDGE: Right. I don’t know if you were involved in an ERP implementation. There’s all kinds of these stories of, yeah, we dropped $5 million last year and it doesn’t work. Well, it does work. You don’t work.
DAVE: You just have to use it right, like any other tool. I love analogies, and you might have picked that. So many times you have people trying to hammer in screws and use a screwdriver to put in a nail. So, yeah, the end result of that is going to be a whole lot of friction and pain.
LEDGE: Split the wood. Absolutely. We jumped off there, but finish some of those stories. You’re so right on the transformation front. The human stuff, I mean, it’s almost like Pareto Principle, is like 80% human. It almost doesn’t matter. The technology is the cost of entry.
It’s so cool that you can do these things. I even noticed, when I looked at the technology before we talked, you're very honest about the fact that there’s a human interface there. To what extent do you care about the mistakes? If it’s an inconsequential fact or whatever, get your best. But if it’s critical business or personal health information or whatever that is, these things really matter. So at what level and how much do you want that human intelligence to intervene and be human?
DAVE: Well, I’ll tell you, there’s a fascinating part of it. I mentioned those aircraft hangers full of paper. A lot of times what those folks are looking for is to try and build models. So you can have 5% error and still develop a pretty decent model, a really good model for what you’re trying to do. You’re looking back over decades. You’re looking at tons and tons of data so you can establish pretty good patterns.
Now, the exact opposite of that is a lot of these operational workflows that we deal with. Let’s think about it. What’s 1 out of 1000? What is that? Like 0.1%? I don’t do math good on podcast, but 0.1%, so you could be at 99.9%.
How do you feel, though, if you’re that 1 out of 1000 and your personal medical report gets mailed to somebody in another state?
LEDGE: Or your claim gets rejected. There’s critical stuff here.
DAVE: Right. Even at 99.9%, you don’t want to be that one.
That’s where all of a sudden this whole thing of, well, single human data entry is like 93% - let’s say it’s 90% probably more realistically. Our software can do 93%, but at the end of the day we were saying that, 99.9% somebody’s data is getting messed up.
At the end of the day, it has to be accurate. That’s where the whole human, that friendly robot mentality, really came into play with what she mentioned.
We have what we call the review interface that allows their team now to focus on much more productive work, and just spend a fraction of the time looking at the digitized data. Even in there, that’s where my team really comes into play – building a lot of these validations, and sometimes even to some extent business logic, that’s fed into the review. Now they’re not even looking at everything. The system is saying, we’re confident on this, we’re not as confident on that. Oh, this is required field on the form, so if there’s nothing there then I’m going to flag it for you. Oh, this needs to be nine digits, I’m going to flag it if it’s not.
Now we’re even narrowing the scope down and they have a lot more time for other productive work.
LEDGE: So it’s adding the intelligence in there, that some human doesn’t have to read all the things to figure out which thing might be wrong.
Finish the story. Our last topic. What replaces that glee of failure? When those light bulbs come on and somebody’s life is better sitting in that job, finish the story on that side, because I have to believe that’s probably highly rewarding for you.
DAVE: When we have these customers in there… I was asked that question in a slightly different way by our former CEO. He was like, “What’s the positive side? I was like, “When it works it’s just magic.” When you’re sitting there with them on a screen share and they’re using the system and this stuff is cranking through. It’s just like this horribly messy handwriting. They’re looking at it side by side and it’s nailing every single thing and they’re able to just click, good, good, good, and the stuff is going through, yeah, it’s like magic.
Like you said, that is an extremely rewarding feeling and they love it. They keep us very busy.
LEDGE: Absolutely. I could see why you’d make some friends. It sells itself after the fact there.
Well, Dave, this is super interesting. Thanks for spending the time. Great insights, and keep up the good work.
DAVE: Sure. Thanks a lot. Thanks again for having me. Really enjoyed getting to meet you and talk.