Latest / Code Story | Startup Podcast for CTOs, CEOs and Technical Founders / S12 E22: The Million-Token Context Mirage: Why RAG is Far From Dead in High-Stakes, Zero-Error Global Tax Compliance with Alex Bowcut, Head of Engineering at Sphere

Transcript

S12 E22: The Million-Token Context Mirage: Why RAG is Far From Dead in High-Stakes, Zero-Error Global Tax Compliance with Alex Bowcut, Head of Engineering at Sphere · Code Story | Startup Podcast for CTOs, CEOs and Technical Founders

• 0:00
  [SPEAKER_01]: In order to build a tax engine, which is essentially adding applicable tax to a transaction, you need to know in-depth each jurisdiction that you support, right?
• 0:10
  [SPEAKER_01]: Some jurisdictions tax.
• 0:12
  [SPEAKER_01]: Certain products, some of them don't tax other products, they have special rates, etc.
• 0:17
  [SPEAKER_01]: and so traditionally the incumbents use these massive teams.
• 0:21
  [SPEAKER_01]: And so it's just a huge human-intensive process, very prone to error, as you might expect.
• 0:26
  [SPEAKER_01]: and it's an inhibitor to growth, it's a big reason why these incumbents didn't expand, and so our takes fierce take and what I'd built when I first joined was what eventually we would call T-R-M, which is the tax review and assessment model.
• 0:40
  [SPEAKER_03]: Hi, I'm Alex Boca, I'm the Head of Engineering at Sphere.
• 0:47
  [SPEAKER_03]: This is Co-Story.
• 0:49
  [SPEAKER_03]: a podcast bringing you interviews with tech visionaries.
• 0:52
  [SPEAKER_03]: Six, six months moonlighting goes.
• 0:53
  [SPEAKER_03]: It's the last and all of the backgrounds who share what it takes to change an industry.
• 0:58
  [SPEAKER_00]: I don't exactly know what to do.
• 1:00
  [SPEAKER_03]: It doesn't go as to get right.
• 1:02
  [SPEAKER_02]: who built the teams that have their bad company is its team's help each other, which is proud of our team.
• 1:09
  [SPEAKER_03]: Keeping scalability top of mind, all that infrastructure was up there.
• 1:13
  [SPEAKER_03]: Yes, we've been fighting it as we grew up.
• 1:14
  [SPEAKER_03]: Total waste of time.
• 1:15
  [SPEAKER_03]: The stories you don't read in the headlines.
• 1:18
  [SPEAKER_02]: It's not an easy thing to achieve.
• 1:19
  [SPEAKER_03]: To get yourself a deficit of off, try to begin to ride the ups and downs of the start-up line.
• 1:26
  [SPEAKER_02]: Need to really want it.
• 1:27
  [SPEAKER_03]: Not just about technology.
• 1:29
  [SPEAKER_03]: All this and more on code story.
• 1:32
  [SPEAKER_03]: time your host's Noel Fart, and today how Alex Boca is building the own AI native techs engine, enabling in-to-end compliance automation across the world.
• 1:49
  [SPEAKER_03]: Alex Boca is from Salt Lake City, Utah, originally, but has been in the Bay Area for the last four years.
• 1:56
  [SPEAKER_03]: He's always been interested in computers and was in middle school when smartphones were blooming.
• 2:01
  [SPEAKER_03]: He took part in jail-breaking gadgets and such and all of these things led to the natural interest in computer science and mathematics, and eventually startups.
• 2:09
  [SPEAKER_03]: Now set attack is married with two Australian shepherds.
• 2:13
  [SPEAKER_03]: He and his wife enjoy
• 2:21
  [SPEAKER_03]: Alex was approached by the founder of his current venture.
• 2:24
  [SPEAKER_03]: He was approached, while he was working at another startup, to tackle the creation of AI control and scaling within the company internally.
• 2:33
  [SPEAKER_03]: When he joined, he immediately started changing the game and creating the initial version of their assessment model, T-RAM.
• 2:43
  [SPEAKER_03]: This is Alex's creation story at Sphere.
• 2:53
  [SPEAKER_01]: Spear is a global revenue-based compliance engine.
• 2:57
  [SPEAKER_01]: So any sort of revenue-based compliance, that's what we're interested in working on.
• 3:02
  [SPEAKER_01]: And the first and most obvious one of those is indirect tax, which is called sales tax in the US and called BATGST internationally.
• 3:11
  [SPEAKER_01]: And so it's spears in the business of supporting as a software, some of the biggest customers, some of the biggest companies I should say.
• 3:18
  [SPEAKER_01]: that are growing in the AI space, folks like lovable, replete, wind serve, 11 labs, all of these folks use sphere.
• 3:26
  [SPEAKER_01]: As a solution for their sales tax needs, so of course companies need to collect sales tax from their customers.
• 3:31
  [SPEAKER_01]: They need to eventually pay that money, they don't get to keep it, unfortunately, eventually they need to pay that money to the tax authorities, and so spheres a solution that makes it very easy to do that, not just in,
• 3:44
  [SPEAKER_01]: the U.S. states, so there's a few different solutions on the market already that I've got some support domestically for the U.S.
• 3:51
  [SPEAKER_01]: But sphere focuses internationally.
• 3:54
  [SPEAKER_01]: So we do support the U.S. as well.
• 3:56
  [SPEAKER_01]: That's kind of table stakes now.
• 3:58
  [SPEAKER_01]: We focus on how can we be a one-stop shop for all of our customers whether they're selling in.
• 4:02
  [SPEAKER_01]: Do you United States, Australia, Japan, Uzbekistan, you name it, we want to support it.
• 4:10
  [SPEAKER_01]: Nick, the founder, Nick Rutter.
• 4:12
  [SPEAKER_01]: He's a single founder and a solo founder.
• 4:16
  [SPEAKER_01]: And he reached out to me right when he was pivoting an old business into what would become his fear.
• 4:22
  [SPEAKER_01]: And he actually, he just found me on LinkedIn.
• 4:24
  [SPEAKER_01]: Small world in San Francisco, small but not so small.
• 4:27
  [SPEAKER_01]: He reached out to me on LinkedIn.
• 4:28
  [SPEAKER_01]: I was at a different start-up at the time working on it.
• 4:31
  [SPEAKER_01]: And it
• 4:32
  [SPEAKER_01]: Nick is a great, he's an incredible founder, and I think I have a pretty decent eye.
• 4:36
  [SPEAKER_01]: Profounders I worked with a number of them and seen a lot of them in action, and from our first conversation I could tell that he had the makeup of what makes a founder great, and I really bought in on this vision that he had, despite it being very early, he had a couple contracted engineers that were working on the proof of concept of what would become sphere.
• 4:56
  [SPEAKER_01]: So it was quite early, but I really bought into the vision of what he saw and the whole and the market internationally that I was kind of all in from the first conversation.
• 5:08
  [SPEAKER_03]: So this will be interesting.
• 5:09
  [SPEAKER_03]: I typically ask about, you know, the MVP for a product, right?
• 5:14
  [SPEAKER_03]: A typical thing, and you read the founders themselves a lot.
• 5:17
  [SPEAKER_03]: Being head of engineering, what was your MVP when you joined the company?
• 5:20
  [SPEAKER_03]: Was that first project you jumped in and took over, ran with, and, you know, what sort of tools are you using to build it and how long to take you to bring to life?
• 5:31
  [SPEAKER_01]: I was hired originally to focus on our AI features.
• 5:35
  [SPEAKER_01]: So Sphere has a bunch of AI features that we've built that are internal features for our tax researchers.
• 5:42
  [SPEAKER_01]: If we back up just to set the stage for one second on how traditionally the incumbents in our space have gone about solving some of the problems that we've solved with AI.
• 5:52
  [SPEAKER_01]: So, in order to build a tax engine, which is essentially adding applicable tax to a transaction, you need to know in-depth each jurisdiction that you support, right?
• 6:04
  [SPEAKER_01]: Some jurisdictions tax, certain products, some of them don't tax other products, they have special rates, etc.
• 6:11
  [SPEAKER_01]: And so traditionally, the incumbents use these massive teams of human tax researchers, essentially tax lawyers.
• 6:18
  [SPEAKER_01]: to go through the legislation in Alabama, for example, right and understand the legislation and then fill out these big matrices and excel of what's taxable in Alabama, what rate.
• 6:30
  [SPEAKER_01]: And so it's just a huge human intensive process, very prone to error as you might expect.
• 6:36
  [SPEAKER_01]: And it's an inhibitor to growth.
• 6:38
  [SPEAKER_01]: It's a big reason why these incumbents didn't expand.
• 6:41
  [SPEAKER_01]: Internationally, from the beginning, are later on.
• 6:44
  [SPEAKER_01]: It's very difficult to scale these huge teams of warriors as probably isn't a huge surprise.
• 6:49
  [SPEAKER_01]: And so our takes fierce take in what I'd built when I first joined was what eventually we would call T-RM, which is the tax review and assessment model, which essentially
• 7:00
  [SPEAKER_01]: acts as like an analyst under our human tax researchers, which we do still have some human tax researchers.
• 7:07
  [SPEAKER_01]: And it makes it a first pass at these determinations is what they're called.
• 7:12
  [SPEAKER_01]: But a lot is taxable in this jurisdiction.
• 7:14
  [SPEAKER_01]: What's not taxable?
• 7:15
  [SPEAKER_01]: What are the rates?
• 7:16
  [SPEAKER_01]: So the AI makes a first pass at these determinations, and then we allow, you know, the human expert goes in and reviews them for correctness and makes changes where necessary.
• 7:25
  [SPEAKER_01]: So this was the first thing that I came in to build was this first version from conversations that Nick and I had, as I was getting ready to join, and then joining, of how we could enable our human tax researchers, who, at the time, we had one contractor one, how could we enable them to move way more quickly?
• 7:43
  [SPEAKER_01]: and they were moving in other places and also allow them to work in other languages.
• 7:47
  [SPEAKER_01]: For example, right, legislation in Japan is not written in English as an example.
• 7:53
  [SPEAKER_01]: And so this was the first thing that I worked on.
• 7:55
  [SPEAKER_01]: And I think it was a great first project and a great first thing to build up from an MVP perspective because it adds such a clear metric of success.
• 8:05
  [SPEAKER_01]: So the game is all about accuracy, right?
• 8:07
  [SPEAKER_01]: How accurate can this system be?
• 8:15
  [SPEAKER_03]: I'm curious about in that, maybe a decision or trade off you had to make in building that and really establishing the T-RAM, right?
• 8:24
  [SPEAKER_03]: I'm curious about maybe as of around approach, you know, and I hear you saying that it was very clear, it was a very clear metric there, but I'm curious about a decision or trade off you had to make and how you coach with it.
• 8:37
  [SPEAKER_01]: I think maybe the biggest trade-off we had to make at the beginning was a laser focus on a particular type of product, so it's very difficult to, we didn't want to try and boil the ocean and go after physical goods, which gets very complicated with like grocery items for example.
• 8:57
  [SPEAKER_01]: So at the beginning, I think the first thing that we did is we really wanted to focus on digital services, right, electronic services, specifically SaaS.
• 9:05
  [SPEAKER_01]: So I think that was the biggest thing from the beginning.
• 9:07
  [SPEAKER_01]: That was a trade-off that we made.
• 9:09
  [SPEAKER_01]: And that was the decision that we made at the beginning that we wanted to focus on SaaS.
• 9:13
  [SPEAKER_01]: We wanted to get SaaS right, SaaS is notoriously difficult from a sales tax perspective in the US and to some extent internationally.
• 9:21
  [SPEAKER_01]: And so we really wanted to focus on SaaS.
• 9:24
  [SPEAKER_01]: at the beginning, and we broke it down as granular as we could get.
• 9:29
  [SPEAKER_01]: So really focused on a certain type of product, how can we iterate quickly on T-RAM, get feedback from our human experts, and trying to do that across a lot of different products is going to be very difficult.
• 9:42
  [SPEAKER_01]: So I think that was the main trade-off that we took from a business perspective going into T-RAM.
• 9:48
  [SPEAKER_01]: I think from a technical perspective,
• 9:51
  [SPEAKER_01]: there are many trade-offs from an architecture perspective that I think I made out of wanting to move quickly.
• 9:57
  [SPEAKER_01]: So building systems that were very agile and maybe wouldn't scale to abillian determinations or something like this, but we wanted to be able to iterate really quickly, take the feedback from our human experts, put it into the system.
• 10:11
  [SPEAKER_01]: And so those were the technical trade-offs I was making were around speed over trying to build the ultimate scalable system for this or something
• 10:24
  [SPEAKER_03]: So you've got the MVP of that working, how have you progressed and matured that?
• 10:29
  [SPEAKER_03]: And I think to wrap in a box a little bit, and this is kind of in general for the company too.
• 10:32
  [SPEAKER_03]: How do you build road maps?
• 10:34
  [SPEAKER_03]: How do you go about deciding that?
• 10:35
  [SPEAKER_03]: Okay, this is the next most important thing to build, or to address was sphere.
• 10:40
  [SPEAKER_01]: That's a great question, and I think spheres approach here is a big reason for its success that we've had over the last couple of years and could growth trajectory that we'd been on.
• 10:50
  [SPEAKER_01]: We worked very closely with our customers.
• 10:53
  [SPEAKER_01]: We have Slack channels with all of them.
• 10:54
  [SPEAKER_01]: We're very in tune with what they need on the tax side of things, around the revenue compliant side of things.
• 11:01
  [SPEAKER_01]: So the way that we set road maps, the way that we decide what's going to be prioritized, is very much influenced by the customers that we have.
• 11:09
  [SPEAKER_01]: And I think that's served as well from the beginning.
• 11:11
  [SPEAKER_01]: Tax is obviously a very complicated space.
• 11:14
  [SPEAKER_01]: It's obviously regulated and there's certainly very low margin for accuracy.
• 11:20
  [SPEAKER_01]: We need to be very accurate.
• 11:22
  [SPEAKER_01]: And so we can't, like I said, a minute ago, we can't boil the ocean.
• 11:26
  [SPEAKER_01]: So we set these road maps based on what are our customers, our current ones, and maybe some that are in the pipeline that we think are close to maybe.
• 11:34
  [SPEAKER_01]: signing, what do they need?
• 11:36
  [SPEAKER_01]: How can we support them?
• 11:38
  [SPEAKER_01]: And how can we support them in a way that's the way they want to be supported something that's hands off for them?
• 11:43
  [SPEAKER_01]: And so we're in a bit of a spot where part of what drives our roadmap as well are legislations all over the world, right?
• 11:50
  [SPEAKER_01]: There's legislations that meet and they make rules around indirect tax and then we have to be ready agile enough to adjust to that.
• 11:57
  [SPEAKER_01]: Sphere, I think, has a unique road mapping setup where we're considering what our customers want, but we also need to consider there's upcoming regulation changes in Brazil, for example, so can we set things up so that we're ready for those regulation changes in Brazil and there's never ending list of these sorts of things.
• 12:14
  [SPEAKER_01]: So it's an interesting mix of customer needs and geopolitical happenings around the world.
• 12:21
  [SPEAKER_03]: What about team?
• 12:22
  [SPEAKER_03]: I'm curious about team and how you, your team and, you know, how the company approaches building teams and what do you look for in those people that indicate that they are the winning horses to join you?
• 12:34
  [SPEAKER_01]: I think we've been very particular on who we bring on.
• 12:38
  [SPEAKER_01]: We don't have an extremely arduous interview process, but I think despite all the growth and we've been very adamant on keeping our technical bar on the engineering side at kind of the level that we've had it from the beginning, which is we want to hire folks who are highly autonomous, take high ownership over the things that they work on and are very technically gifted.
• 13:01
  [SPEAKER_01]: And so we've stuck to our guns there, I think,
• 13:04
  [SPEAKER_01]: Always, as you're growing, there's some pressure to fill seeds with bodies to help out, but we've been very particular about who we bring on.
• 13:12
  [SPEAKER_01]: I think sphere as a particular culture of customer obsession, of product obsession, of working hard, frankly, and we expect a lot from our employees.
• 13:23
  [SPEAKER_01]: I think we give back a lot as well.
• 13:25
  [SPEAKER_01]: But we've been particular, we think, we're a small team for the size, the revenue that we're doing into stage that we're at, we're a small team, not just on the engineering side, but across the company.
• 13:36
  [SPEAKER_01]: And that's been very deliberate.
• 13:37
  [SPEAKER_01]: We think that in this day and age, like the day and age of AI.
• 13:41
  [SPEAKER_01]: There's much more you can get out of the same size of people compared to 5, 10 years ago, and I think we really embraced that.
• 13:48
  [SPEAKER_01]: That being said, we are, we still hire aggressively and we're constantly looking at folks, but we really want folks who care about sphere, care about the company, and especially at a startup, you need people who are willing to do what it takes to try and help the company succeed.
• 14:04
  [SPEAKER_01]: I think there's a culture of no task as beneath anyone at the company, whether that's
• 14:09
  [SPEAKER_01]: Nick, the founder, or me, is head of engineering.
• 14:12
  [SPEAKER_01]: We're happy to do whatever it takes for the company to succeed, and we're not worried about titles or a hierarchy or anything like that.
• 14:23
  [SPEAKER_03]: Let's move into scalability.
• 14:25
  [SPEAKER_03]: And this will be really interesting.
• 14:26
  [SPEAKER_03]: This is in your space and your bread and butter.
• 14:29
  [SPEAKER_03]: How has scale been approached at sphere?
• 14:32
  [SPEAKER_03]: And I'm also curious, you know, have there been interesting areas and interesting stories where you've had to fight scale as you've grown?
• 14:39
  [SPEAKER_01]: I think scale for us maybe means a couple of different things, especially in engineering.
• 14:43
  [SPEAKER_01]: So one part of scale would be scale of requests coming in to our API.
• 14:50
  [SPEAKER_01]: So we integrate with lots of billing systems, for example, Stripe.
• 14:55
  [SPEAKER_01]: So we're a first-party integration with Stripe, essentially we're a drop-in replacement for their own pack solution.
• 15:02
  [SPEAKER_01]: So you can very easily switch from Stripe Tax to using Sphere.
• 15:06
  [SPEAKER_01]: There's only a couple companies in the world that have this sort of thing.
• 15:09
  [SPEAKER_01]: And that's been great for the business, obviously, but it also means that Stripe holds us to very high SLAs.
• 15:15
  [SPEAKER_01]: Essentially, they hold us to the same SLAs they hold themselves, which they aim for five plus nines of availability,
• 15:25
  [SPEAKER_01]: Understandable, they're a big company and they should be hitting those sorts of things for a start-up.
• 15:30
  [SPEAKER_01]: I think that's a big, that's a big ask and it's been a big ask and we've had to work quite hard to hit those sorts of targets, but that's something that when we think about scalability, when we're thinking about our APIs, we're thinking about
• 15:43
  [SPEAKER_01]: SLAs, right?
• 15:45
  [SPEAKER_01]: We just can't have any downtime.
• 15:47
  [SPEAKER_01]: A accompanying that comes the number of requests that are coming in.
• 15:51
  [SPEAKER_01]: So as we're scaling the business, there's tens of thousands more requests per minute that are coming in.
• 15:57
  [SPEAKER_01]: Every time we sign a big customer who's doing a lot of transactions, all of those transactions are going to result in multiple API requests to our API.
• 16:06
  [SPEAKER_01]: So we're thinking about scale from that perspective.
• 16:09
  [SPEAKER_01]: And I think we've done a good job.
• 16:11
  [SPEAKER_01]: from the beginning of choosing technological solutions that are able to scale with us in that.
• 16:17
  [SPEAKER_01]: So we're using Kubernetes.
• 16:19
  [SPEAKER_01]: We can spin up pods as we need them, and then on the other part of scale, I think for us that would look like data volume.
• 16:27
  [SPEAKER_01]: So we need to sink in all of our customers' transaction data, all of their historic transaction data, actually.
• 16:34
  [SPEAKER_01]: That's obviously a huge amount of transaction data.
• 16:37
  [SPEAKER_01]: As we bring on customers, there's tens of millions, hundreds of millions of transactions.
• 16:41
  [SPEAKER_01]: They need to be synced into our system.
• 16:44
  [SPEAKER_01]: And we need to do that this quickly as possible.
• 16:46
  [SPEAKER_01]: One of the ways that sphere kind of sells itself above some of the competitors is how quick our onboarding process is.
• 16:52
  [SPEAKER_01]: and part of onboarding, of course, is like sinking in the data that we need in order to finish the onboarding.
• 16:58
  [SPEAKER_01]: And so we're thinking about scale from a database perspective as well, like a data volume perspective.
• 17:03
  [SPEAKER_01]: How do we deal in process all this data that we need to be sinking in in as short of time as possible?
• 17:10
  [SPEAKER_01]: And on the database side, we've been using AWS Aurora Postgres, and I think that served us very well.
• 17:16
  [SPEAKER_01]: I think it was a wise decision when we switched over to using it pretty early on and has really scaled with the system.
• 17:22
  [SPEAKER_01]: I think at some point we'll probably outgrow it, but for now it's been a fantastic solution.
• 17:27
  [SPEAKER_01]: So I think we've been pretty forward looking.
• 17:30
  [SPEAKER_01]: And it's somewhat easy to forecast, like we know our API requests are going to the number of them per second is going to go up significantly over the next six months.
• 17:40
  [SPEAKER_01]: So we're taking actions today to make sure that we're in a spot to support it.
• 17:43
  [SPEAKER_01]: Same thing on the data volume side.
• 17:48
  [SPEAKER_03]: Okay, Alex, as you step out on the balcony, you look across all that you've built this for at sphere.
• 17:53
  [SPEAKER_03]: What do you most proud of?
• 17:55
  [SPEAKER_01]: One thing for sure that I'm very proud of is T-RAM, which we've talked about.
• 18:00
  [SPEAKER_01]: One of the most gratifying parts of the job, so far that really make the long nights or whatever worth it is, when we bring new human tax researchers on to our team, to work with T-RAM.
• 18:14
  [SPEAKER_01]: Without fail, they have the same reaction as they start to use it.
• 18:18
  [SPEAKER_01]: They're like, oh my god, this is so amazing.
• 18:21
  [SPEAKER_01]: This, you know, I wish I would have had this 10 years ago.
• 18:24
  [SPEAKER_01]: It would have saved me so much time.
• 18:25
  [SPEAKER_01]: You all should sell licenses to this because I know 100 people from my network that would pay so much money to use something like this.
• 18:33
  [SPEAKER_01]: So this has been every time we get that feedback.
• 18:36
  [SPEAKER_01]: It's just you get the warm fuzzies inside and it makes you feel like you're doing something.
• 18:40
  [SPEAKER_01]: You're really on to something that's really moving the needle on a process that is slow and archaic in other places, so I think that's been a really enjoyable part of it.
• 18:48
  [SPEAKER_01]: I think another really enjoyable part has been the engineering team that we've put together at Sphere, I think.
• 18:55
  [SPEAKER_01]: All of them are extremely talented, I think this is definitely the most talent-dense team I've ever been a part of.
• 19:01
  [SPEAKER_01]: And despite being so technically gifted, it's been
• 19:07
  [SPEAKER_01]: Work together as a group and really push yourself as a group to accomplish more than you thought was possible.
• 19:14
  [SPEAKER_01]: And I think if you talked to each of the engineers on our team, they would echo that.
• 19:18
  [SPEAKER_01]: They thought they were possible for them to accomplish a certain amount before they joined Sphere, and they've really been pushed beyond what they thought they could do.
• 19:27
  [SPEAKER_01]: And it's really quite the experience to kind of go through that as a group and see everyone kind of grow and be able to take on more and do really incredible things, frankly, for us, for all the team.
• 19:38
  [SPEAKER_01]: As we have, we're still around only 10 engineers.
• 19:41
  [SPEAKER_01]: to be supporting and building the sorts of things that we are, that's hugely gratifying, I think, especially for someone like me sitting where I am.
• 19:52
  [SPEAKER_03]: Okay, let's flip the script a little bit, Alex.
• 19:54
  [SPEAKER_03]: Tell me about a mistake you made, and how you and your team responded to it.
• 19:58
  [SPEAKER_01]: This is a very tough industry.
• 20:00
  [SPEAKER_01]: I mentioned before, it's always a moving target.
• 20:04
  [SPEAKER_01]: Things can change based on legislation, things that are totally outside of your control.
• 20:09
  [SPEAKER_01]: And I think we've certainly made mistakes from an engineering perspective thinking about maybe like data models or the abstractions we're thinking about these problems.
• 20:18
  [SPEAKER_01]: Those turned out to not be right because we didn't foresee that Singapore was gonna have these certain rules that kind of don't fit into the box.
• 20:27
  [SPEAKER_01]: that we had built based on our abstractions or data models.
• 20:30
  [SPEAKER_01]: And so we need to backtrack a little bit.
• 20:32
  [SPEAKER_01]: Or there's unexpected rules in Alabama for how something is taxed that doesn't fit cleanly into the abstractions that we made.
• 20:40
  [SPEAKER_01]: And I think that those have been the major things where I think we get better at it every time we kind of run into one of these new cases that we haven't faced before.
• 20:49
  [SPEAKER_01]: And we build a more extensible modular system.
• 20:52
  [SPEAKER_01]: But those have been the main mistakes, I think that we've seen are,
• 20:55
  [SPEAKER_01]: It's not around like choices of technology or anything like that.
• 20:58
  [SPEAKER_01]: It's more around us as engineers.
• 21:01
  [SPEAKER_01]: We don't have a part, none of us are tax people, right?
• 21:04
  [SPEAKER_01]: I started its fear and I didn't know anything about sales tax other than you see it on your checkout session when you're buying something.
• 21:11
  [SPEAKER_01]: So we're all learning as we go and trying to learn from the human tax experts, and it's a very difficult space.
• 21:18
  [SPEAKER_01]: There's a reason that no one has ever modeled but global solution to the sales tax indirect tax problem.
• 21:24
  [SPEAKER_01]: And so there's been plenty of mistakes we've made along the way that if I think back,
• 21:29
  [SPEAKER_01]: of some of the decisions we were making 12 months ago, I wish I could go back in time and kind of grab myself and be like, no, you haven't considered how this is going to work in Switzerland, and that's kind of, that's going to break the abstraction.
• 21:41
  [SPEAKER_01]: So I think those have been the main mistakes, I think.
• 21:47
  [SPEAKER_03]: So let's move forward then.
• 21:49
  [SPEAKER_03]: This will be exciting.
• 21:50
  [SPEAKER_03]: I know you joined this team because you're excited about what's being built.
• 21:53
  [SPEAKER_03]: You've created some amazing things.
• 21:55
  [SPEAKER_03]: Tell me about what the future looks like for a spear.
• 21:58
  [SPEAKER_03]: For the product for where the industry is going for your team company, all the things.
• 22:04
  [SPEAKER_01]: I think this is a very exciting time to be in the industry that we're in.
• 22:09
  [SPEAKER_01]: Going back to an earlier question about what's been gratifying about this, I think that's been another thing that's been very gratifying is seeing changes around the world happening that make it even more important for businesses to be using a solution like sphere.
• 22:23
  [SPEAKER_01]: Makes you feel like you're really on the right track here.
• 22:25
  [SPEAKER_01]: So an example of that would be Ian Voicing.
• 22:28
  [SPEAKER_01]: This is something we're adding support for.
• 22:30
  [SPEAKER_01]: It's a new product feature for us.
• 22:32
  [SPEAKER_01]: E-invoicing is essentially, there's many governments around the world where they're mandating that the government systems are essentially in the loop of transactions that are happening.
• 22:43
  [SPEAKER_01]: So when you want to send an invoice to a customer, that needs to be sent through an E-invoicing system, which is like a government system.
• 22:50
  [SPEAKER_01]: Because the governments want real-time reporting of all the transactions that are happening in their jurisdiction.
• 22:55
  [SPEAKER_01]: So this is like a burgeoning field is just beginning and we sit in a really good position.
• 23:01
  [SPEAKER_01]: We're already connected to folks building systems.
• 23:03
  [SPEAKER_01]: Ian voicing is very, you need to know the tax, so tax is very important for it as well.
• 23:08
  [SPEAKER_01]: So adding support for Ian voicing is something that we're already working on rolling out, but we'll take some time to support across all of the countries.
• 23:17
  [SPEAKER_01]: We support from the sales tax side.
• 23:19
  [SPEAKER_01]: We also have our eye on tariffs, for example.
• 23:22
  [SPEAKER_01]: You have all of our customer's products, QCADLog data, and tariffs are perform of tax.
• 23:29
  [SPEAKER_01]: That tax depends on the type of good, and that is very similar, of course, to the way sales tax operates.
• 23:35
  [SPEAKER_01]: And so we're looking to support folks from the terrorist perspective and then maybe the one other feature set that we're looking at right now as well is what's called input tax.
• 23:46
  [SPEAKER_01]: So internationally, you can reclaim some of the tax you pay on your purchases as a business, which of course is very interesting to our customers because it lowers your tax bill.
• 23:56
  [SPEAKER_01]: So it's money that goes back in their pocket.
• 23:59
  [SPEAKER_01]: So they're very interested in making sure they get their input tax right, and that's this ultimate play that we're looking at is really outside of just sales tax, even though even on the sales tax side, there's no global solution just for sales tax before sphere, but we're not resting on our laurels there.
• 24:17
  [SPEAKER_01]: We have much broader plans.
• 24:19
  [SPEAKER_01]: We have bigger terms that we want to eat into.
• 24:22
  [SPEAKER_01]: We'd like to eat significantly into the
• 24:28
  [SPEAKER_01]: And that's today, people pay them for things like e-in voicing an input tax date.
• 24:32
  [SPEAKER_01]: They pay a lot of money to these firms and we think a lot of it could be a software solution.
• 24:37
  [SPEAKER_01]: And we think sphere is perfectly placed to take advantage of these changing regulations and serve our customers and make it as easy to sell this tagline that we repeat sometimes.
• 24:49
  [SPEAKER_01]: We want to make the world one market.
• 24:51
  [SPEAKER_01]: It should be as easy to sell in California as it is to sell.
• 24:55
  [SPEAKER_01]: In Germany or Nigeria, it shouldn't matter for you as the owner of a business you should be able to just sell your products to whoever wants to buy them.
• 25:05
  [SPEAKER_03]: Let's switch to you, Alex.
• 25:07
  [SPEAKER_03]: Who influences the way that you were?
• 25:09
  [SPEAKER_03]: Name a person or many persons are something you look up to and why.
• 25:14
  [SPEAKER_01]: There's been a lot of people in my career that have shaped the way that I work, the way that I approach problems, whether those are technical or personal problems like building the team out, personnel problems.
• 25:27
  [SPEAKER_01]: But I think if I think back to the beginning of my career, I worked at a company called MoZenda at the beginning of my career, and I worked for someone there named Caleb.
• 25:38
  [SPEAKER_01]: Like anyone at the beginning of my career, I didn't understand how things worked.
• 25:41
  [SPEAKER_01]: I wasn't sure how to be a great employee.
• 25:43
  [SPEAKER_01]: I wasn't sure how to make a big impact on a team, and I think he really enabled.
• 25:48
  [SPEAKER_01]: me to understand those sorts of things.
• 25:51
  [SPEAKER_01]: How do I sniff out what's most important for the business, right?
• 25:55
  [SPEAKER_01]: What's going to have the highest impact?
• 25:57
  [SPEAKER_01]: How do I do the things and build the teams that are going to help hit the goals that the business is made?
• 26:03
  [SPEAKER_01]: And so thinking back, I think Caleb is really up there.
• 26:06
  [SPEAKER_01]: In fact, a business that I founded later on, I founded it with him.
• 26:09
  [SPEAKER_01]: So I think he's had a really
• 26:12
  [SPEAKER_01]: effect on me and then jumping forward maybe.
• 26:15
  [SPEAKER_01]: I've been working with Nick the founder of sphere for a couple years now and I think he's had a very big impact as well.
• 26:22
  [SPEAKER_01]: I think Nick is very focused.
• 26:25
  [SPEAKER_01]: He's relentless in the way that he approaches problem.
• 26:28
  [SPEAKER_01]: So I think I've never seen someone quite like Nick before in the way that they are.
• 26:33
  [SPEAKER_01]: see a problem and then just attack it, eat it into submission so to speak, just what's the problem?
• 26:39
  [SPEAKER_01]: How do we solve it?
• 26:40
  [SPEAKER_01]: How do we move forward?
• 26:41
  [SPEAKER_01]: What's good for our customers?
• 26:42
  [SPEAKER_01]: If I think back two years before I joined Sphere, I think that was a muscle I didn't have nearly as much and I think that's a really powerful thing to be able to do is to be able to.
• 26:52
  [SPEAKER_01]: see a problem for what it truly is, make a plan for how we're going to solve it, and then not relentlessly attack that problem until it's no longer a problem, it's actually a string.
• 27:03
  [SPEAKER_01]: And I think that honestly is the story of sphere is we see our weaknesses, and we turn those into strings, and we've done that over and over again, and if you keep doing it, then in any case you're going to build an incredible business, and I think that's what we've been doing.
• 27:21
  [SPEAKER_03]: Okay, Alex, last question, so you're getting on a plane and you're sitting next to a young entrepreneur who's built the next big thing.
• 27:27
  [SPEAKER_03]: They're jazzed about it, they can't we just show it off to the world and can we show it off to you right there on the plane?
• 27:31
  [SPEAKER_03]: What advice do you give that person having gone down this road a bit?
• 27:35
  [SPEAKER_01]: I think my main point of advice would be to chase the dream, continue to chase the dream that it brought them to that point.
• 27:43
  [SPEAKER_01]: Obviously building a product.
• 27:45
  [SPEAKER_01]: especially building like the MVP of a product is one thing but then building a business around it is another thing and I think sometimes it's easy for founders to lose track of what was the original dream that they had when they set out and started building what was the real point of the product or what was what was it really that motivated them to do this incredibly hard thing which is building something out of nothing.
• 28:10
  [SPEAKER_01]: Looking back on my career, that would be the main advice I could give is just to never forget what was that original thing that pushed you to do it and how do you keep serving that instead of serving other things that might be trying to steal your attention or your time or your focus or pull the product in ways that you hadn't thought.
• 28:30
  [SPEAKER_01]: How do you stay true to the original vision?
• 28:32
  [SPEAKER_01]: All the successes and startups, or in my career, I've seen it's always been remaining true to cut core prints of what you set off to do.
• 28:42
  [SPEAKER_03]: I think it's fantastic advice.
• 28:43
  [SPEAKER_03]: Well, it's thank you for being on the show today and thank you for telling your creation story at Sphere.
• 28:49
  [SPEAKER_03]: Yeah, thank you, Noah.
• 28:52
  [SPEAKER_03]: And this concludes another chapter of Coat Story.
• 29:03
  [SPEAKER_03]: Code Story is hosted and produced by Noah Labhart.
• 29:06
  [SPEAKER_03]: Be sure to subscribe on Apple Podcasts, Spotify or the podcasting app at your choice.
• 29:12
  [SPEAKER_03]: And when you get a chance, leave us a review.
• 29:14
  [SPEAKER_03]: Both things help us out tremendously.
• 29:19
  [SPEAKER_03]: And thanks again for listening.