KYC 43 | Helping Your Customers

 

Slater Victoroff never thought he would grow up to be an entrepreneur, even though he was exactly that from an early age, selling individual Dum Dums that he had bought in bulk to his friends (with a nice markup). Fast forward to today, and Slater is the former CEO-turned current CTO of Indico Data. With a focus on helping their customers do their job in a more fulfilling way, Indico processes unstructured data such as audio files, contracts, invoices, applications, etc.

Join us for a conversation that takes a turn into really deep thinking around humans and data.

Watch the episode here:

Listen to the podcast here:


About Slater Victoroff

Slater Victoroff is the Founder and CTO of Indico, an enterprise AI solution for unstructured content that emphasizes document understanding. He has been building machine learning solutions for startups, governments, and Fortune 100 companies for the past seven years and is a frequent speaker at AI conferences.

Indico’s framework requires 1000x less data than traditional machine learning techniques, and they regularly beat the likes of AWS, Google, Microsoft, and IBM in head-to-head bake-offs. Indico recently announced a $22 million Series B raise.

Slater has been building AI, machine learning and deep learning solutions for the enterprise for the better part of the past decade, having worked with everyone from the federal government to two-person startups to the Fortune 100. He has educated hundreds of business users on successfully implementing deep learning through a simple framework that helps executives rapidly accelerate the adoption of the technology in their businesses.

Slater attended Olin College of Engineering before pursuing Indico full-time after its acceptance into the Techstars Boston Program.

Helping Your Customers Do Their Jobs In A More Fulfilling Way With Slater Victoroff

Collaboration Among Humans, Data, And Automation

I am excited to have Slater on the show. I loved our conversation. It went down paths that we didn’t anticipate. It’s powerful for the C-Suite to read this one because we’re talking about things that make a difference in the business. His company is focused on data mining, unstructured data documents, invoices, contracts, automating processes, all of that stuff. The level of conversation we had is far deeper than that. It’s about how do you build a culture that can create products and services that matter? How do you get the right customer in the room?

He tells us stories about how he started out with one customer set that because it’s who they were and then had to migrate over time to a new customer set that were the users of their software. It’s a fascinating story and the depth that we go into around that is something that everyone in the C-Suite needs to read because you might be thinking you’re going down one path and you’ve got to switch gears if you want to grow your product, your service offerings and your customer base, you may have to switch who your customers are.

Slater Victoroff is the CTO of Indico Data that’s based in Boston, Massachusetts. What I enjoyed about this conversation is he is the CTO, an engineer, a left-brain guy and he brings such a human element into his work, the conversation and he’s a super deep thinker. You can tell it’s an interesting conversation. I pulled out a lot of things. The one thing that I enjoyed going deep into is when he referenced that he took a class in User-Oriented Collaborative Design. Be sure and tap into that piece of this conversation because it was interesting. With that, we’ll dive in and jump into our conversation with Slater Victoroff.

Slater, welcome to the show. We’re so happy to have you here.

Thanks so much for having me. I’m thrilled to be here.

Let’s start at the very beginning. Tell us about how you got to where you are now, backing up as far as you want to, whatever you want to share with us. Even as far back as to where you went to school, where you grew up, those kinds of things and your journey to get to where you are now.

I was thinking a little bit about this coming into the show. A lot of folks are familiar, maybe with the key parts of the story. I did not think of myself as an entrepreneur at all until after my first year at school. I worked at startup within Pearson. I got bit by the bug and then fell very much in love with machine learning and the rest is history. One thing that I haven’t necessarily talked about before that is a little bit of a funny story. It goes back way further than this to what I think was probably my first real business that I ran with a friend.

I realized that the big bags of Dum Dums you could buy from the corner store at CVS were way cheaper than you would think they were. It was $2 or something for this massive bag of Dum Dums. I was like, “Kids at my school will pay more than that for Dum Dums.” I set up this whole business buying candy from the corner store in middle school and then bringing it to school. I had a friend, he is a VC, he was excited and he started doing the books. It was fun.

Yet you don’t consider yourself an entrepreneur at that age? That’s very entrepreneurial.

I didn’t think about it at all. I don’t think I’ve mentioned it until now.

That’s a great story. I love it. There is nothing like a product with a good market potential.

Middle schoolers with lollipops. That’s a great market.

Good product-market fit there. You don’t have to do research. You know it.

What happened after middle school?

There is always going to be a need for humans to consume data. And there’s always going to be a reason for machines to interpret. Click To Tweet

After middle school, ups and downs through high school and whatnot but where things took off for me, I did a lot of extracurricular that I enjoyed. Shout out to Science Bowl, Science Olympiad and Quiz Bowl. Those were a few academic competitions that I loved doing through high school. I volunteered with a few of those throughout undergrad and some of them even to the current day. Ocean’s Bowl as well. I should make a specific shout out, too. They have competitions that run out of MIT locally. Undergrad, Olin was a very interesting place. The key thing to call out there is that Olin believes in this notion of project-based education.

For those that aren’t familiar with Olin, it’s a very small engineering school and it is just engineering. It’s right near Wellesley. When I say small, it’s about 80 students per class and it’s undergraduate only. There is no graduate curriculum, no graduate school there whatsoever. It’s engineering only, so it’s very similar to Harvey Mudd in that way.

One of the ways in which they’re interesting in addition to the project-based focus, which says that every class is going to be oriented around building something real rather than tests. There are tests, but few and far between, it’s primarily building something at the end of the class as close to that as you can get. They also force every engineer there to get effectively a design minor and they believe very much in teaching engineers the why of what they’re building. One of the jokes is that Olin’s Oliners very often go on to be product managers rather than engineers. I don’t think that’s necessarily a bad thing but certainly, there is also a lot of very strong engineers coming out of Olin.

Was that focus on experiential learning the reason you attended Olin?

Absolutely. I had never seen anything like that growing up. I knew that that was a separate mode of education; I believed all education was this very strict lecture-based test-taking mode. I was pretty decent at that. I wasn’t good at the homework, but the test-taking and the lecture-following, I was pretty good at that part of things. It opened my eyes. Going there and participating in a lot of these do-first experiments. They know how different this mode is because their admissions process is also very different.

They invite candidates to school for a weekend to get that flavor of Olin. For me, that was transformative. The funny part of the story is that after I went to candidates, I withdrew my application to every other school and Olin then rejected me. I had to go back and plead on my hands and knees and then I got on the waitlist. A year later, I got admission.

It’s a testament to persistence. You didn’t throw up your hands and say, “I’m not going to Olin.”

It wasn’t the smartest thing at the time but I still knew that it was what had to happen.

Fast forward to your role, tell us about your company, the problem you solve, the customers you work with, how you got to this point in your career.

What we’re building is enterprise software. It’s very much oriented at the subject matter expert. This is someone who is processing unstructured data like contracts, text, images, audio, things along those lines. When you have any human process that involves interpreting large amounts of that information, that’s where Indico plugs in. We plug in in a way where we’re giving the subject matter experts, that’s the person processing the invoice, contract or loan application. We’re giving them power over a tool that will help them to automate the key repetitive aspects of their work. Both do their job in a better, more consistent way and personally fulfilling way.

How did you come across this focus for your company? How did you determine that was a problem that needed to be solved?

It took a while. We had a couple of missed swings as well. That’s always what happens. You take an idea out and the second you start bringing in front of customers, you start realizing you’re doing the wrong thing entirely. The initial goal and direction that we wanted to head in was we wanted to make it accessible. We started with making that accessible to developers and we eventually realized that there was this amazing opportunity to make it accessible to non-technical individuals. We saw that there was a big need for this technology within the enterprise.

We specifically found some key application areas around automation and analytics and new application building even around unstructured data. We didn’t have that at the start. That all formed gradually with time over the course of 6 to 7 years as we developed this and talked with customers and brought them things that started maybe as a handful of tools and like, “We’re going to walk together with you.” That vision emerged with time.

It sounds like you were listening to your customers right out of the gate, which is a very near and dear topic to my heart as well as Tony’s. Who are your customers and tell us about how you build a culture of relationship with your customers within your organization.

KYC 43 | Helping Your Customers

Helping Your Customers: The art of working on a team and coming together to work towards an end goal even when there are personal differences is incredibly hard. It’s one of the most important skills and you rarely get that taught at all.

 

We focus on the Fortune 500. This is large enterprises in the United States, maybe a little bit beyond in North America and maybe a little bit into Europe as well but the United States primarily. BFSI, that’s the key vertical. We certainly work in other verticals but that’s where the bulk of our customers are and that’s banking, financial services, and insurance. There is a smattering of use cases across that but the question of how you infuse that user-centric mentality and it’s very interesting at Indico in particular because we are dealing with very bleeding edge technology. That’s a technology that is so advanced that the adoption of that technology creates some active pain.

That often is seen as being at particular odds with the customer experience. The way that we think about it is that by focusing on that end-customer experience, we focus on the power that gives us to experiment more broadly with the technology that is powering these experiences. One of the other things that we think is powerful is that new technology, we think is interesting in so far as it impacts the user experience. Every time we see a new paper come out, that’s what we’re asking is, “If we can now load a model in 30 seconds instead of 30 minutes, what new things is that going to allow us to do?”

That’s a good way of getting everyone around the company excited about this notion of focusing on the impact of the user because you get this cool nonlinear payoff where you can find something that’s like, “I didn’t think this was impactful change. I thought this was some little architectural tweak.” This opens up a whole world of possibility.

You said it’s been 6 to 7 years you were bringing tools to the market to different users. You started with the developers and you’ve moved to this more user-centric approach. We’ve had some other entrepreneurs on the show and it’s always interesting to talk with them about those first couple of years, that early journey and who were you listening to. I get this idea from other conversations we’ve had. It’s like, “I’m listening to myself. I’m seeing a cool idea but then I start listening to someone else.” Who was it you were listening to in that early journey that you got into the developers and then what was that a-ha moment that you said, “We’re focused on the wrong people here?” Or was it an a-ha moment?

First off, why did we focus on developers in the first place? Because we were developers. Very easy there. We’re like, “We know what we want, so let’s build what we want.” Very simple. While we executed on that, we were definitely a bit closed off to the broader world. That’s how I would reflect things. What I would, frankly, characterize things as is the feedback that ultimately pushed us to focus less on the developer or at the very least to bring other users importantly into the fold as first-class citizens were the developers themselves.

It was interesting when you talk to a lot of these early users, we had this notion as college students that developers are working on cool, fun projects on the side and that’s what they do. Now at an enterprise, at a place like MetLife, that is not what they do. They have end users and these customers that they’re trying to satisfy. We came to this realization like, “We’re giving them these tools. They don’t understand how to build the thing and we don’t understand how to build the thing.” We’ve got to build the thing.

We’ve gotten to the point where we can build the end product. We can provide that to people and because we’ve still got that rich developer core, there are all these APIs and configurability that you get because they are still an important part of the equation. We can say, “We do solve this end problem for this non-technical user.” We make those original users and that broader stakeholder network happy in a holistic way.

I’ve worked closely with developers between BI, CRM and all the things over the last 20 years. I was their customer early in my career and then I became the consultant. I’ve played those roles. I do find it interesting, too because you were saying, “How they can do their job in a more personally fulfilling way?” I want to call that up because that’s a significant piece. Any company that’s going to succeed, especially in tech, has to be figuring out how do we help our users have a more fulfilling life? It’s not what you probably learned in school. Maybe it was, I didn’t learn that in school. What’s your take there?

Most people don’t learn that in school. It was one of the things I was lucky to learn at Olin. There are some grad schools that do have equivalent classes but I’ve never heard of anything similar to this being taught at the undergrad level. In your second semester, sophomore year, you take a class called User-Oriented Collaborative Design. You think about that title for a second and you’re like, “That’s very interesting.” That’s what the entire course is. It’s very interesting because it’s an engineering course. It’s at an engineering school. You’re not building anything. The entire course from end to end, the idea is to figure out what to build and convince us that it is a good thing to build.

You importantly do it for a user group that you don’t understand that’s a little bit further outside your zone. That is what’s interesting is you learn through that experience. There is such a non-correlation between the complexity and how cool you think it is and how the users love it. One of the most eye-opening moments to me in that whole process and a lot of engineers get this impression as well, is sometimes in that early design process, you have to put together terrible mock-ups. It’s like, “This is barely functional.” For us, it was a little bendy straw with a piece of yarn that went through it, and then there was a piece of cardboard with Velcro on it.

That’s the whole thing. You can draw with the yarn. It gets stuck to the Velcro. You can imagine what, these lo-fi prototypes look like elsewhere. When you show that to a user and they love it. They’re like, “You don’t have to build anything. I don’t care. I want this. It’s awful but I don’t care. This is what I want.” That’s a transformative moment for engineers. You can go your whole career without having a moment like that, if you’re not plugged into the design process.

I hope there are some business professors in curriculum development people reading this episode because it strikes me why in business school are there not these teaching people how to be collaborative. What a huge education because that doesn’t always come naturally to people. It’s also that sometimes that you don’t know what you don’t know thing about how important it is to be collaborative. I love that. That needs to go beyond engineering school.

Learning to design with other people is so hard. There is a whole science to how they put the groups together. There is a concept of silver bullets where you’re allowed to veto team members because it’s so sensitive. It’s such a raw and emotional experience. They won’t talk to each other for years afterwards. That’s rare but it happens.

It’s an education.

By focusing on that end customer experience, we focus on the power that gives us to experiment more broadly with technology that is powering these experiences. Click To Tweet

There was this interesting study. If folks know the thing with the pasta and the marshmallow is where you’re supposed to either build a bridge that supports the most weight or build something that’s as tall as possible. The way they did this, the variation that I thought was interesting is that they put people in groups of their colleagues. First, they did them as solo and they had a business major, liberal arts major and engineering major. In the solo competition, maybe the business person won and maybe the engineer was the worst. When they had people work with groups of their peers, they had this inverted effect. Engineers are often taught to collaborate with each other effectively on engineering things.

They worked well. The business majors are not taught to do this. They’re taught that life is a competition. They’re at each other’s throats. They can’t build anything. They do even worse than an individual business major when you put them all together. It’s interesting because I completely agree is that the art of working on a team and all coming together to work towards some common goal, even when there are personal differences, it’s incredibly hard. It’s one of the most important skills and you rarely get that taught at all in school.

I had heard an expression, people talk about soft skills and someone I know, he has a podcast as well and he was like, “We don’t refer to them as soft skills. We refer to them as essential skills.” Learning how to collaborate, especially with your customers, if you’re trying to solve an issue mutually, learning that collaboration skill is an essential skill. I love that.

One other that I’ve highlighted from my personal perspective is writing, especially in the modern age, people underestimate how important writing is.

I’m so glad to hear you say that, especially coming from a left-brain type field. I’m so happy to hear you say that. My degree is in journalism. I know I have a bias toward that but when I see a piece of bad writing come across, it immediately makes me distrust them. Not personally but the quality of their work or their ability to communicate. It’s a pretty instant disqualifier.

I feel like the average quality of prose has dropped. I’m sure you understand this. I didn’t realize what ridiculous timetables journalists have to work on. When you’re turning a 500-word piece in three hours, it’s impossible to produce something of quality on that timetable in my book, especially that everything is increasingly digital. A lot of people don’t necessarily understand how difficult it is to convey ideas clearly in text and how incredibly useful that skill is.

Especially with customers because if you can’t communicate with clarity and this is a lesson I learned early on is being prescriptive without being pushy. Writing in a sense that conveys leadership versus that line between collaboration and leadership and you have to take that right balance between, “Here is what we recommend you to do and here is how we can help,” versus, “I don’t know, what do you want to do?”

There is a subtle line almost between management and leadership as well that is hard to appreciate in some ways. Depending on the customer’s direction one or the other might be correct.

We’re way off course here. We told you it was going to be a conversation but we’ve got some other questions to come back to. What is fascinating is this whole idea of diving into collaboration, teaching collaboration, leveraging that you need that when you work with your customers, you need that when you work with your team internally. Any good project has to have a requirements document, has to have not just what we’re doing, but here is why we’re doing it. It’s interesting to communicate that clearly so that someone who wasn’t in the room can understand it.

This is something Betsy and I have talked about there is that chair in the room. When it is all the engineers around the table, you’ve got an empty chair for the customer in the room and you’re like, “Are we meeting their needs and what they said or not?” It is fascinating to think about that. We can talk and we can share ideas but if no one documents that in writing, by the time we get back together or we get back to our desks, we might have a different impression of what that is.

The other thing that’s interesting is there are often many different layers at which that needs to be captured. Something that was a big learning for us early on is understanding what is in the specification. Who is this for? What detail is it appropriate for us to nail down at what point in the process? How far along in the process are we? Are we talking about building this next week or in two months? There are different levels of specification. Negotiating that and helping both engineers understand, “We’re putting together a marketing requirements document to get on the same, very high level of what we’re going towards.” Understand that there is utility in having them in that conversation.

Conversely, when engineering is getting into the middle of like, “We’re putting a pen to paper and we’re going to be coding this next week. We need to go through in a lot of detail and make sure we’ve got these right.” Also, making sure the business understands that they still have to make some decisions at that point. It’s this two-way street because it can be awful if you get either one of those out of whack because they do impact each other no matter strongly you try to abstract these business functions away.

It’s not just for entrepreneurs, whether you are employed or leading a company of your own, the ability to collaborate and communicate it’s such a lesson and particularly as it relates to customers. People receive information differently, too so you have to be in tune. Some people like it in writing, need the face-to-face or shoot me a text with the bullet points. You have to understand how people receive that information as well.

That’s a lot of why crafting that requirements document in an appropriate piece, understanding, “We need the abstract. We need these bullet points. There are a couple of figures that we need to add to explain these key points.” It’s an important part in and of itself understanding how we are going to craft this and then put it together in a way that everyone feels good about the process.

KYC 43 | Helping Your Customers

Helping Your Customers: If you give AI something that’s not clean to emulate, it will do it, but then your output won’t be clean and then you haven’t really solved your problem.

 

We have some clients that when we hop on a Zoom call, which is the way of life now. We hop on a Zoom call, it’s like, “Cut to the chase. What do we got to do? OK. Moving on.” It’s great and that’s fine. We have other ones that are, “How was your weekend? What are you doing this weekend? What are your holiday plans?” All that stuff and being tuned in to that piece of what makes your customer tick and getting to know your customers at that level, too, is an important skillset.

I went to this public speaking co-curricular, a side thing in undergrad. I remember there was a great question that the instructor asked and he says, “What profession has the best public speakers across the board?” Either of you have any guesses?

I don’t.

I’m going to guess sales.

Reasonable guess but not all salespeople. A reasonable salesperson but there are some more relationship-type people. Some folks get deep. Overall, but not positively. You might also think of politicians. That’s another one where you certainly usually associate it with them but there are alternatives. You can’t say every single politician must be a good public speaker.

My wheels are spinning. It’s probably not medical professionals.

It’s an interesting guess. Stand-up comedians, it’s a very interesting point. It’s a lot of that face-to-face piece. Understanding how do you read an audience? When do you stretch this joke out? No matter what, if you’re going to be even a halfway decent comedian, you have to be a master of that timing and audience impression. It is something that’s also important for Zoom to that point, having something to react with your audience around.

Being able to read a room and shift on the fly is a phenomenal skill.

It’s key.

I was on a call with a CEO. They’re still remote. He was talking about communication on his team and he’s like, “How do we do this?” It was interesting because he said, “We have got so many channels of communication. I’ve had to make these arbitrary rules like you don’t send documents in Slack, you have to send those via email and you communicate this via text.” He admitted at the end, he’s like, “Maybe it makes life easier for me. I know I’ve seen stuff. I don’t even know where to look for it anymore because we’ve got all these channels of communication.”

I’m going to segue us back into what Indico does because as he’s talking about this, I’m sitting here thinking about all the different data and information that’s out there and what we lose in the process. A lot of what your company does and the software that you have, it helps people grab that estimated 85% of unstructured data communications, invoices, contracts and that stuff. There is all this communication that’s coming toward us. How do we bring it together and leverage it, so we get the most out of it?

One of the questions that I sometimes get that’s interesting along these lines and it’s almost philosophical in a way because people will sometimes ask me the question, “Aren’t documents going away? What happens when documents go away?” They’ve got this notion like, “There is a world where everything is web formed and every single process is this perfectly manicured structured input.”

I don’t think that is such a thing but the question comes down to, “What is a document? What is unstructured data?” Why do we use this stuff when I think everyone can agree there are these massive downsides? It doesn’t have all of the nice transparency and ease of analytics that we want with our data sources.

What it comes down to, in my view, is that data is meant for human consumption. That’s the thing, why we write documents, because that is how humans want to consume data. There is always going to be a need for humans to consume data. There is always going to be a need for machines to then also interpret at least in an ancillary way that same data made for human consumption.

How do you build a culture that can create products and services that really matter? Click To Tweet

That’s what it comes down to in a lot of ways is that humans have this very unstructured, messy, fuzzy way of talking with each other. It’s not that there aren’t rules, directions and consistency to it, we all see it in the middle of it but it’s this very difficult contextual problem. You have to understand what’s happening. There is no problem like NLP to humble you to what’s going on in the human mind.

One of my daughters has an exam and she is studying rhetoric. I had to go back and refresh myself as I was helping her study the rules of rhetoric and all of that. I’m like, “I hated this when I was in school.” I’ve been in the text mining world for almost 20 years. I know the value of it but then I realized also there is this reality of people don’t communicate this way. How do you balance this? There are rules and structures that we’re supposed to live by but we don’t live by them. It’s a challenge.

You see it going all the way back. I was thinking about Plato’s famous like humans are bipedal whatever birds with toenails and no wings, whatever he said that we were. We constantly trying to do this. We constantly try to winnow and complete and formalize. As quickly as we tried to create these normalizations, we realized this was not quite accurate. One of the big a-ha moments was the advent of Twitter because when you talk about something that put a dent in the entire NLP world, it’s the advent of Twitter.

You had all of these techniques, work well for formal texts because until 1999, 2000, I don’t know when Twitter came about, you’ve all got this very formal text. That’s what you’re going. That’s the entire world of analytics. Suddenly, Twitter becomes incredibly interesting and important. It doesn’t ask this question of, “How do we expand out to a new domain?” it’s, “How do we throw everything we know about language out the window?” No other place do you see such a profound breakdown in the formalism of communication and yet no human needs to be coached on how to understand Twitter. It comes so naturally to us.

Narrowing down how you want to say it in originally 140 characters and having that parameter around it. It’s interesting.

It forces you to cram tons of nuance into these little snippets, which again, it was compounding problems for traditional approaches.

Speaking of hearing and listening, if you were sitting in a coffee shop and there were two peer executive-level people talking to each other, what would you want to overhear them say about your company?

That we’re awesome and we have great engineers and that they’re going to pay us a lot of money to buy our products.

I’m pretty sure that’s what every entrepreneur would say.

Some of this I already do here, so I don’t know how aspirational this is supposed to be, but I do have a lot of pride in our team, not just the engineering team. A special pride in the engineering team, being a CTO, but the Indico team overall. We’ve got such an amazing, heartfelt group of folks. We’re incredibly conscientious around the user’s needs. We’re very supportive of each other. It’s great. Unfortunately, we have started to get a bit of recognition. Our engineers are very highly sought-after. Thankfully, they’ve got good spam filters, too.

That’s good news/bad news thing when your people become so good that they’re in demand.

It makes it clear to the folks at Indico that they are valued and learning good stuff and it forces us to make sure we treat them right. It’s a win-win at the end of the day.

We’ve covered a lot of ground here. I love the conversation. This has been awesome. There is one question that’s been on my mind. For people that are in the space where we are in the customer data mining space and you said your company’s at the bleeding edge of it. When we look at AI and machine learning, all of the other techniques that are out there, what’s the limit that companies should be thinking about in customer data? It’s a heavy topic, I know, but it’s one that is important for CEOs and the C-Suite to be aware of because things are changing so fast.

Are you thinking particularly in terms of third-party data sharing and stuff like that or more in terms of large-scale rollouts of automation?

KYC 43 | Helping Your Customers

Helping Your Customers: In no other place do you see such a profound breakdown in the formalism of communication, and yet no human needs to be coached on how to understand Twitter.

 

I think both are important.

Where do you want to start?

Let’s start with the large-scale rollout first.

There are 2 to 3 main issues we see customers run into depending on their maturity level or what they’ve done from an automation perspective already. One issue that we see a lot of customers in, we call it the accuracy treadmill or they get stuck in the muck of their own processes. Especially a lot of traditional template-based solutions, reject-based solutions, the total cost of ownership or the cost to constantly update that is so incredibly high that not only can you only afford to go after a very small number of use cases and very few even fit that template, but also the expense of maintenance over time is a mess. It makes it prohibitive to go after in a scale that way.

That one piece is being cognizant that different use cases across your organization are not necessarily going to be solved by the same technology. There are very few technologies out there that can solve your full gamut into these cases. The second big issue that we see is the services bottleneck. This is the most obvious way that especially a lot of vendors try to patch that gap is when you have the notion that every single use case you do is going to take a very large data science services motion, whether that’s an internal data science group that’s doing novel model development or an external services group that’s doing AI for you.

It doesn’t matter. It takes a long time. Very expensive. It means out of the gate that you’re looking at months to years before you’re getting these models into production just because development has certain laws of physics that apply to it.

The final one, and this shows up after you’ve solved the first two bottlenecks. The first two bottlenecks that first one, you’re only going to be able to get after 2 or 3 use cases in that way. The second bottleneck, you can probably get after ten use cases running in that services mode but no way 100. This third bottleneck is what you’re going to run into, call it dozen of use cases as you start trying to get to 100.

That’s scaling the formalization of knowledge and processes. What we often see as the big gap is that you don’t want to be optimizing your process while you’re trying to automate it or apply AI to it. AI, these techniques, they’re very good mimicry machines but if you don’t give them something pretty clean to emulate. If you give them something that’s not clean to emulate, they will do it but then your output won’t be clean and then you haven’t solved your problem. Often, the formalization and getting enough subject matter expert time to teach machines effectively becomes the bottleneck building that internal rigor.

This has been a fascinating conversation. We so appreciate you joining us. Unfortunately, time has flown by. One final question for you, on our show, we like to give our guests an opportunity to give a shout out to a non-profit or a charitable organization or somebody that’s doing good in the world. Is there an organization you want to tell us about?

Two. One is People Making a Difference locally and the other is the TEALS Organization. People Making a Difference runs the local Ocean’s Bowl and helps with some of the other Science Bowl and other local competitions. TEALS has a co-teaching model where they have software developers teach at high schools to help high schools build out computer science curriculums.

Thank you so much. Tony, any final questions?

I have lots more questions but we ran out of time. Thank you, Slater. We might have to have you come back again. I appreciate it.

It was a total pleasure. It was a lot of fun.

Thanks so much, Slater.

That was the most in-depth conversation I’ve had on a philosophical level this week. I loved it. There are a couple of key points that I wanted to pull out in case the audience may have read it and it passed by. One is the focus of his company is to help the users do their job in a more personally fulfilling way. What a why behind the work that they do. That stood out to me as someone who understands. He is of that generation, I don’t know how old he is exactly, but he is a late-Millennial or early Gen-Z, probably late-Millennial, but it’s that whole idea of personal fulfillment, doing something good in the world for other people is what drives them.

Why do we use documents? It’s data meant for human consumption. Click To Tweet

It seems like the culture of his company follows on that. The other piece that stands out to me is the conversation we had around data for human consumption. What’s the definition of a document? These may seem like esoteric questions or philosophical questions, but they’re tremendously important as we’re getting into virtualization, we’re getting into this world where we’re in a space of virtual reality. We’ve got to question, “What does this mean?”

A document used to be a piece of paper, parchment or something like that. Now it means something completely different. Understanding that and thinking about that at a deep level, all C-Suite executives have to do that because if you don’t, your business is going to be facing a startup like Indico. You may be thinking in a very different field than they’re thinking. They’re going to eat your lunch.

What’s striking me as I process the conversation that we had, he is so bleeding edge, technology, where it’s going, machine learning and all of these innovative things and yet he is still very grounded in the foundations of the stuff that were important to us growing up, which is effective writing, communications, the things that are timeless and that still matter despite how fast the world is moving. I love that he brought it back to the fact that writing is such an important skill, even when you’re going down these crazy paths of technology. I appreciated his comments on that. Another great episode, Tony. Thanks, everybody for joining us on the show.

Important Links: