The B2B Podcast Index
Index
All categories
MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
MethodologySubmit
Best of:MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
An independent project byFame
SearchBest episodesGuestsInsightsMethodologySubmit a podcast
Index/Finance/"Tech Finance" with Sasha Orloff
"Tech Finance" with Sasha Orloff artwork

E69: Agentifying the $7 Trillion Tax Payment Network with Solon Angel of Remitian

"Tech Finance" with Sasha Orloff · 2026-06-03 · 48 min

0:00--:--

Key moments - from our scoring

Substance score

55 / 100

Five dimensions, 20 points each

Insight Density11 / 20
Originality9 / 20
Guest Caliber13 / 20
Specificity & Evidence13 / 20
Conversational Craft9 / 20

The tax payment infrastructure powering $7 trillion in annual payments operates on systems designed in 1994, before the internet, e-commerce, or modern fintech existed. Solon Angel spent 15 years building AI for accounting (founding MindBridge, which was acquired by KPMG) before identifying a critical gap hiding in plain sight: once tax returns are filed, the actual movement of money remains manual, error-prone, and rife with fraud. Accountants handle pre-signed checks vulnerable to theft, bookkeepers steal from trust accounts, and high-net-worth clients must manually route payments across 3,000 separate tax endpoints (state, federal, local, various tax forms) that change mid-season without notice. Remitian uses AI agents to handle the compliance complexity - these are not payment processors like Stripe, but compliance-native agents trained on the rules and variations across all 3,000 endpoints. The market opportunity is massive: late fees and penalties alone represent enormous leakage, and an executive order in 2025 banned checks for tax payments. Angel argues this is fundamentally about domain expertise: Silicon Valley historically dismissed accounting startups as unsexy, but VCs investing in the space now are finally recognizing that deep domain knowledge beats surface-level thinking. For B2B operators, CFOs, accountants, and fintech builders serving professional services, this episode explains why payments infrastructure matters more than filing software and how agentic AI solves the compliance coordination problem at scale.

Key takeaways

  • →Tax payment infrastructure has not evolved since 1994; across North America there are 3,000 tax endpoints (federal, state, local, varying by tax type and form) that accountants must manually route payments to, often pre-signing checks that remain vulnerable to theft and fraud.
  • →Accountants and bookkeepers handling trust accounts create massive AML and security risks, which is why an executive order in September 2025 banned checks for tax payments - the problem was recognized as systemic.
  • →AI agents (not payment processors) are required because tax payments are compliance events, not simple transactions; the agentic approach handles rule variation, endpoint routing, and deadline management across all 3,000 endpoints.
  • →MindBridge took 15 years to be recognized because Silicon Valley VCs structurally underestimate high-domain-expertise markets like accounting, biotech, and legal, preferring surface-level analysis and betting on platforms instead.
  • →Remitian's vision is to eliminate tax season deadlines entirely, making tax payments seamless, automatic, and painless through AI coordination - collapsing what is today a fragmented, high-friction manual process into a background system.

In this episode

  1. 1The Broken Tax Payment Infrastructure: Why Checks and Manual Processes Persist in 2026
  2. 2The Origins of AI in Accounting: How DeepMind and Machine Learning Inspired MindBridge
  3. 3Why Silicon Valley Missed the Accounting Market: Domain Expertise and Market Structure
  4. 4From MindBridge to Remitian: Recognizing the Payments Layer Problem
  5. 5The Complexity of Tax Compliance: Why 3,000 Tax Endpoints Require Agentic AI

Mentioned

RemitianSolon AngelSasha OrloffMindBridgePuzzleCaseWareDeepMindOpenAINotioniWinBasisinstead.com

Guests

Solon Angel

Topics in this episode

AI agentsACH paymentsCasewareTax compliance infrastructureAnti-money laundering (AML)RemitianMindBridgeTrust accountsTax endpointsExecutive orders on tax payments

Questions this episode answers

Why do accountants and bookkeepers still handle pre-signed checks for tax payments in 2026?

The tax payment system was built in 1994 before e-commerce existed, and it was never designed for convenience or automation. Accountants inherited the role by assumption, and checks were tolerated by regulators as long as accountants had trust account rights. However, an executive order in September 2025 banned checks for tax payments after recognizing the fraud risk.

How many different tax payment endpoints exist in North America?

There are 3,000 tax endpoints across North America, including federal, state, local, municipal, and county authorities, each with different tax forms and ACH accounts. This complexity means banks and accounting firms must manually maintain lists and route payments correctly, with authorities changing details mid-season without notice.

What is the difference between Remitian and a payment processor like Stripe?

Remitian operates as compliance-event infrastructure, not a payment processor. Tax payments require licensed compliance handling across 3,000 different endpoints with varying rules; they cannot be simplified into bipartite payment models. AI agents trained on compliance rules are necessary to route and execute payments correctly at scale.

Why did it take 15 years for venture capital to invest in AI for accounting?

Silicon Valley VCs structurally underestimated accounting as a market due to surface-level thinking about domain expertise. They dismissed startups like MindBridge, not recognizing that deep professional services expertise creates defensible modes. Only when baby boomer partners began retiring and accounting firms opened to outside capital did VCs recognize the $7 trillion opportunity.

What is Solon Angel's vision for tax payments five years from now?

Angel wants to eliminate tax season deadlines entirely. The tax payment experience should be seamless and painless, with AI agents automatically handling all compliance, routing, and timing so taxpayers no longer face a stressful filing season with hard deadlines.

What our scoring noted

Our reviewer’s read on each dimension, with quotes from the episode.

Insight Density

11 / 20

There are genuinely non-obvious facts scattered throughout - the 1994 creation of the tax payment system, 3,000 tax endpoints in North America, 12% misclassification rate, the cartel overpayment-refund laundering scheme - but these are interspersed with lengthy anecdotes, a sponsor break, a Golden Gate Bridge metaphor, and startup-narrative padding that dilutes the signal-to-noise ratio considerably.

tax payments are not payments, they're compliance events. And that's why you cannot have a bipartite compliance level
12% of tax payments are ah, misclassified by the internal systems and there's humans under tax receivership trying to quickly within two weeks reassign them properly before but the taxpayer still occurs two weeks of penalty

Originality

9 / 20

The reframing of tax payments as compliance events (not generic payments) and the vision of eliminating tax deadlines by spreading payments over time are genuinely fresh angles, but the episode leans heavily on recycled startup tropes - small team beats bureaucracy, Tesla vs GM, Warren Buffett 'stick to what you know,' 'never underestimate a small team' - that undercut the originality.

I want to kill the deadline of tax season. Uh, it's so stressful. It's stupid.
The reason why Tesla is Tesla and General Motors did not create a Tesla is a prime example of that

Guest Caliber

13 / 20

Solon Angel is a genuine domain practitioner - early AI in accounting via Mindbridge, five years at Caseware building global distribution, angel investor with LP relationships, and board roles at accounting firms - making him credibly experienced rather than a pure thought-leader circuit guest, though Remittian is still early-stage and the depth of at-scale validation is limited.

I spent five years in CaseWare, which is the third largest provider of accounting software in the world
I run into a DeepMind um, in 2014 in Paris where they explained how deep reinforced learning mechanisms will allow to self adapt to repetitive video games

Specificity & Evidence

13 / 20

The episode delivers a solid layer of concrete specifics - named client firms, real dollar figures, named regulations, and operational metrics - that ground the discussion meaningfully, though some numbers (e.g., $85B penalties) are asserted without sourcing and several claims about the upcoming partnership are deliberately kept vague.

85 billion of late fees and penalties per year in North America. It's a tax payment network of $7 trillion in the US alone
we cover 300 of the 3,000 tax endpoints in North America

Conversational Craft

9 / 20

The host occasionally asks productive follow-ups - pushing on the general counsel story, probing why the problem hasn't been solved, and challenging the 'AI company hiring more people' contradiction - but he defaults to affirmative filler ('I love it, I love it'), inserts a long personal anecdote about the Golden Gate Bridge that adds nothing, and rarely challenges the guest's bolder claims with meaningful scrutiny.

Can you share a little bit more about this thing you were talking about with the general counsel where they're like, you're like no, you actually, you can do this
I love it, I love it

Conversation analysis

Computed from the transcript - who did the talking, and the words that came up most.

Share of words spoken

  • Speaker A74%
  • Speaker B26%

Most-used words

accounting30clients30money18firms18market17payments16different15accountant15accountants15first15today14checks14payment14back14world13mindbridge12

Episode notes

Sasha Orloff sits down with Solon Angel, CEO of Remitian, to explore why tax payments remain one of fintech's most overlooked infrastructure problems. They discuss the outdated systems still powering tax compliance, how AI agents are enabling a new payments layer for accountants and taxpayers, and why the convergence of regulatory change, fraud prevention, and agentic AI could transform the $7 trillion tax payment ecosystem into a seamless, deadline-free experience. - SPONSORS: Notion Boost your startup with Notion - the ultimate connected workspace trusted by thousands worldwide! From engineering specs to onboarding and fundraising, Notion keeps your team organized and efficient. For a limited time, get 6 months of Notion AI FREE to supercharge your workflow. Claim your offer now at ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Puzzle No more waiting for month's end. No more manual reconciliations, just financials that keep up with you. Puzzle is leading a new movement with an AI first ledger, built for speed, automation, and real time decision making.

Full transcript

48 min

Transcribed and scored by The B2B Podcast Index.

Speaker A: Behind the scene. What happens is that people until recently still today rely a lot on checks that they will drop in their content pre signed for their content to mail in. Last year one of the offices that got a burglary and there was hundreds of checks white pre signed that got stolen.

Speaker B: No. Why is this still happening in 2026?

Speaker A: Well, there was an executive order by this administration on September 1st of last year that says no more checks for tax payments. So before that there was a certain tolerance to checks. But checks are a great way to commit fraud. Massive fraud. Cartels launder money through tax authorities sometimes like it's crazy.

Speaker B: If Remedian succeeds in the way that you envision it, what do you think the tax payment experience looks like five years from now?

Speaker A: Seamless. Painless. And there's no more deadlines. I want to kill the deadline of tax season. Uh, it's so stressful. It's stupid.

Speaker B: Welcome to Tech Finance. I'm your host, Sascha Orloff, repeat founder and current CEO of Puzzle and self professed finance nerd. On this show we dive into the pulse of B2B fintech. From newly funded startups transforming the office of the CFO to major industry shifts, deals and partnerships that are reshaping how we think about finance. You'll hear the most important developments, essential analysis, and of course, AI innovations. This show is brought to you by Puzzle and the Turpentine podcast network. Let's dive in. All right, welcome everybody. Today's guest is someone who has spent 15 years salon. That's. That's quite a long time. We're going to dive into this building AI for accounting and financial infrastructure. And this was long before most people had sort of thought about AI in the accounting world on a day to day basis. And I think what's going to make this conversation super fascinating is that his company, Remittian, is attacking a problem that has been hiding in plain sight in front of all of us. Once a tax return is ready to file, the actual movement of money is still pretty manual. You guys all know this. Bank porters, bank portals, mail checks, like what jurisdiction is going where, and all of this money in just late fees and penalties alone. And so while the filing software market we've all heard of is like a big, big industry, the actual payment layer is broken. And Remidian believes in AI. Native infrastructure is finally going to fix that. So, Salon, welcome to the show. Uh, excited to have you here.

Speaker A: Thank you for having me. Seychelle.

Speaker B: Cool. So you previously founded Mindbridge, which is one of the Earliest AI companies in accounting. And you said you grew this from a couple thousand dollars in savings into a company worth hundreds of millions of dollars. You've invested in dozens of accounting startups and you're now building what you believe is the missing payments rail for tax and compliance. So tell us, like paint the picture. What is really happening behind the scenes today?

Speaker A: So what happens behind the scenes is both structural and um, assumptions that are pushed on taxpayers. So structurally, actually, Quick quiz. Guess when the electronic file tax payment system was created.

Speaker B: Well, I don't know. Uh, let's go with, uh, you're not, I'm supposed to be interviewing you. What's going on over here? Let's say, uh, ah, 97.

Speaker A: Yeah, close. 94. Oh. So there was not many people using the Internet back then. And um, you know, Fintech was not really a thing. Um, PayPal was not even really launched. And that's when the system that we use today was created. So you can imagine that structurally here there's already a lot of things missing off the bat. Right. It's not a system that was made for convenience. It's not a system made for E commerce. It's not a system made for software companies or Internet company, let alone to use. It's very, very much a banking system in collaboration with a processor done by the government. And the other assumption is, a big assumption is that taxpayers can just do it themselves. As if taxpayers are not busy enough with their lives. Because it doesn't matter, it's an obligation. I mean they didn't catch Al, uh, Capone for murder. They caught him because he didn't pay his taxes. So ultimately everyone has to pay their taxes. Um, and so it was an assumption. It doesn't matter how clunky it is or not. Right. Plus the difference is back in the day, back in 1994, there was less local taxes. So you didn't have county, state, municipal, you didn't have that. Layers of different taxes all using different systems and some not having a system at all. They just give you an ach number and off you go, you're on your own. So behind the scene, what happens is that people, until recently, still today rely a lot on checks that they will drop in the accountant pre signed for the accountant to mail in.

Speaker B: No.

Speaker A: Yes. And so last year one of our clients, which, uh, I'm not going to say who, they had one of the offices that got a burglary and there was hundreds of checks white pre signed that got stolen and the thieves went to buy stuff in town with It. So the situation is quite fascinating to me because accountants are people that bear an incredible workload and suffer in silence and taxpayers have no way around it. They have to pay. Otherwise they hit with late fees and penalties or jail. So that's why we're in this situation where you have very funny, bizarre practices that have been tolerated for a long time. And by the way, I did not discover that on my own. It's an accountant dad was frustrated of having his clients paying late fees and not able to help them. We said I need to do a better care for my clients and to provide a better service. And he went down that rabbit hole of contracting one of my friends to build the first version. And he was right. He had the foresight to see that there must be a better way. Something that is more touchless for taxpayers.

Speaker B: So what you're sort of saying as you're painting this picture, uh, is the payments infrastructure was in the 90s, uh, and the payments was somewhat simple. Uh, back then there wasn't this sort of proliferation of maybe all the state and county e commerce, all the rest types of fees. Is that kind of the view of why is this still happening in 2026?

Speaker A: Well, I think it's still happening for a couple of reasons. But there was also another thing is until last year accountants, like lawyers had the right to have trust accounts. So there was a certain tolerance for anti money laundering authorities to let accountants be in the flow of funds. And also there was less security back then, um, concerns of ID theft. So people would share their banking passcode to their accountant to when they were too busy to do the quarterly installments or when they had lost structures. Look, one of our large CPA firm has a client asked 270Einstein. He has properties across the US he has several businesses, he has family trust, he's a wealthy successful person and he had bookkeepers doing that for him. One of the bookkeepers stole money. He doesn't trust bookkeepers to do that anymore. What does he do? Turns to the accountant and the accountants charge a white glove service that until recently everyone trusted them. But if you look at the case with BDO and Jay Z, I don't know if you heard about it, it was one of the staff of the accounting firm who stole millions of dollars from Fat Joe and Jay Z. So you can't trust humans. Even the employees of a large top 8 VA firm can be trusted with money. So there's a lot of issues around the concept around being compliant with anti money laundering laws that have evolved a lot and at the same time trusting your tax advisors, which is not supposed to be handling money to do that. So that's how we end up in that situation. It's just decades of overlooking the problem and the expectations. If you go to university and you ask students, what does a CPA firm does? The majority of them say they pay people's taxes. So the majority CPA firm, if you talk to them, they say, oh, we don't handle payments. We just prepare the tax delivery return. They send the PDF to the client and off they go because they don't want to have too much responsibility. So there's a gap, there's a discontinuity in the marketplace between the client's expectations and what the structures of the market expect Today.

Speaker B: I was, uh, with my son. We were going on a hike up in Marin, and we were coming across the Golden Gate Bridge, and there was a ton of traffic. And, uh, and he was like, why? What's going on? We're like, on a Saturday afternoon, plenty of traffic. And so on the, on the Golden Gate Bridge, it was just built in a different time for a different era. And there wasn't made for the amount of traffic and volume and types of cars that go across it. And so there's literally a kind of a, a special machine that moves the lanes. Yeah, Twice a day and it drives through and like. And we were like, right behind it as it was like, doing this thing. And so you just see this, like, how do we adapt into a world that at that time they probably thought, oh my God, there's lanes. For the indefinite future to come. Now we're bottlenecked by, uh, having to physically move the barrier of how many lanes exist in any given time. I say that only to set up as sometimes there's these things that live just right in front of us that, that we don't see. So I want to go back and talk a little bit about your first company. You've been doing AI and accounting for long before there was commercially available AI. So what was it that sort of gave you that insight 15 years ago?

Speaker A: DeepMind. So before going back and being an entrepreneur, I spent five years in CaseWare, which is the third largest provider of accounting software in the world, um, after Thompson Rotters and Walter Clowers. And at Caseware, I was exposed around the world. I helped them build a global distribution channel, travel a lot. And there was a lot of issues on the product with the traditional analytical packages that auditors used, uh, to do their work. A lot of training required. It was cumbersome. And if you look at what an audit is, it's very repetitive every year there's very strong standards. And then I quit case where I went to help turn around another startup in AI where we started using machine learning. We used to call that just machine learning and computer vision in physical security presence. And then I quit and then I run into a DeepMind um, in 2014 in Paris where they explained how deep reinforced learning mechanisms will allow to self adapt to repetitive video games and learn quick by themselves. And it just clicked. I'm like I need to be the first one to bring that technology to audit. And yes it was raw and yes it was not very available. There was no open source framework for it. We had to do a lot of things ourselves. We had to build some of our own mutating algorithm we were exploring with OpenAI's 2.5 release even we used some of it to do classification of accounts very successfully. We had a patent on it. But that's what happened to me. It was that. And by the way, that video is still on YouTube today. DeepMind in Paris 2014 it just clicked just like that's the way of the future for audit and I need to be the first one to do it.

Speaker B: The story still is similar, like what's a repetitive problem, complicated, requires all this language. Um, and so there's this narrative that's been oh my God, software is going to replace accountants, machine learning is going to replace accountants, AI is going to replace accountants. Uh, yet for 15 years you've been committed to using machine learning and AI to help solve different problems. So what was the point of view back then? Similar thing.

Speaker A: So the name of the company, MindBridge to bridge the gap between the human and artificial mind, always made an assumption that licensed professionals like doctors, CPAs, if they do so back in the day the average auditors did 16 audit per year. And my view was simple. If one day an auditor does 1,600 audits or 16 at the end of the day they see someone needs to sign the audit opinion and review it. That is a licensed trustworthy humans. That said that system was configured properly. And right now I'm a shareholder of one company called instead.com and instead.com this tax season claimed that a tax ah, Accountant can do 2000 returns with their system with AI but you still need a license. It's a license CPA in the loop at some point. So it doesn't matter if one day we will have 20,000 audits done by one human controller that controls the air traffic control of all those AI systems. At the end of the day you still need to augment the human capacity and professional judgment. So that was my view. There will be always, even if there's one auditor left in the world, there will always be one human left somewhere,

Speaker B: probably a lot more, but I get it. So given you've been doing this for a while, you know, now we see almost all of the major venture funds investing in accounting. Who would have thought right? Sequoia, uh, Andreessen Horowitz and Benchmark Sell and Softbank and Google Ventures and go and you know, et cetera, et cetera, the list goes on and on and on. What do you think changed that all of a sudden? Made accounting and AI such a hot market.

Speaker A: By the way, full disclosure, I'm on the board of two accounting firms, one in the top 20. I have um, excellent relationships with managing partners, um, in that space for a long time. Some of them were my mentor when I was a case ware and I have an incredible amount of respect for what they go through. But I think the accounting profession was the last non professional professional services running business like the accountants. Uh we're running the CPA firm like cash based business that were like landlords collecting rent on the partnership share. Right on the partnership, uh, profit sharing. And so basically there was a lot of dentistry roll ups, there was a lot of roll ups of Laundromats. There's a lot of roll ups on a lot of industry. But there was one industry that never was properly run was accounting services. And the combination of a lot of managing uh, partners, retiring baby boomers retiring in throve in the last 10 years started opening the door to okay maybe we'll let capital from outside benefit from that juicy market. And I think that there's a demographic aspect that we can't forget. But also I think there's a failure from Silicon Valley from acknowledging very high domain expertise area. I'll give you two examples. Tobi from Shopify goes to Silicon Valley to raise money and they did not understand E commerce. They said the TAM of E commerce sites at the time was 60,000 E commerce sites on Yahoo sites. So thus there's not enough market for Shopify. Shopify created a million plus E commerce site on their own, right? So Silicon Valley and VCs have that very structured surface level thinking about deep domain expertise that hurts them. The same thing with Mindbridge. I got kicked out of Litespeed and there's many vseek. When I went to talk about Mindbridge, one of Them very famous, not gonna say the name says you will never convince a big four to invest in new startup. They are so stuck up versus buying new technologies they'll build their own. MindBridge has KPMG Global as a customer and I explained them how I will because I did it case aware before and they dismiss completely because Silicon Valley admitted failure to sell to accountants. They invested in big um CPA style um next gen neo accounting startup that didn't really pick up in the last 10 years. And I think when I read the manifesto um services as a new software. I took it as a capitulation from Silicon Valley to say we're not able to sell tools to those guys because we don't fully understand them. So we should just go after the book of business directly and that's a great thing. But I think Silicon Valley sometimes see so many pitches that they cannot go in high domain expertise area properly. Same with biotech and others. And that's why other investors uh, and other GEOs like for example New York, Chicago, that has m majority of large accounting firms understand that better.

Speaker B: Hey, we'll get back to the conversation in a moment after a word from our sponsors. All right. One of my favorite tools that I use and we use is Notion. Notion is used by thousands of startups around the world as their connected workspace for everything from engineering specs to new hire onboarding to fundraising. Notion has a special offer for listeners. Six months of Ah, free notion with unlimited AI value of over $6,000. You can access the limitless power of AI right inside of Notion to build and scale your company and team with one tool. Get six months free of notion with unlimited AI along with custom templates when you apply for notion.com startups. Once again, that's notion.com startups. P U Z Z L E. Go for it. Building a company is hard enough. You shouldn't have to second guess your numbers too. One day an investor might ask you a question that you just can't answer. The next you realize your books aren't as clean as you thought. Sound familiar? Puzzle is rethinking accounting from the ground up with an AI first layer ledger that eliminates repetitive tasks and gives you real time financial clarity so you can focus on running your business, not fixing spreadsheets. Accounting should work the way you do. Fast, intuitive and built for the future. Come check us out at uh, Puzzle IO. All right, well that leads into uh, something that you, you mentioned before. You don't like playing another man's game. What does that mean when you're thinking about when you're thinking about your businesses.

Speaker A: Look, most startups fail and it's very hard to compete. And I've acquired a network and expertise for 20 years and 15 years building AI companies in the accounting market. Some people are big in Japan. Um, I was striving to be big in accounting. Nobody cared about it at the time, but I invested in a lot of successful startups. I win being one of them, um, or basis and others. And if you look at it, it's very simple. I was top 100 influencer in accounting today and I was the only angel investor in that list. I had the best deal flow on my inbox, right? Like it was incredible. Um, I stopped investing. So to me it's very simple if you are made to. I used to play competitive judo. I went to rugby because it's physical similar. But I'm not going to become a swimmer or ballerina dancer or. You know what I mean? Like when you have a certain skill set and I said, you just have to stay in your game, don't try to compete in another one. I'm not going to become an NBA player, right. I'm going to be a rugby man, right? So it's a Warren Buffett rule as well, you know, stick to what you know and you'll be fine. So that's why I stayed in accounting. And frankly, they're lovely people to work with. A lot of them have insane workloads and they have a true care for their client and they went in that profession out of the purest sense of it, which is of service and doing the right thing. They are a very high, um, integrity profession. So it's quite an honor of a lifetime to see them going through those transitions and see them adopting AI. And sure they give me a hard time when I used to talk about big data in 2008, about AI in 2012, but now when I tell them about some things, they listen a bit more. After two or three times ignoring me.

Speaker B: Uh, I have to say it is such a wonderful group of, uh, people building software for, for accountants. I had the opportunity to work with Mitch, uh, one of the founders of Basis back in the early days and just see his ambition and hunger as he sort of built his own company. And Ellen, from aiwin to here in San Francisco, was generous with her time and grabbed coffee and talked a little about the industry. In fact, our conversations as well. It's a really good group of people because there's so much opportunity that we can all win. We can help the firms we can help accountants in all sorts of different ways. Uh, it's such a big market that's been, like you said, kind of untapped in the, in the traditional VC world. All right, so talk a little bit now about how you kind of turn this insight with Mindbridge into now, what are those influences that, that helped you see the opportunity and decide you wanted to do this full time with Remedian?

Speaker A: So actually it was before that was helping Aywin. Um, I had a lot of problems. So at mayanbridge, we had limited resources and limited time to do a good job. So we had almost like so many use cases came to us over time, but we chose to stick to our lane. The iWin use case in my collaboration with iWin, is because one of the top customers of MyBridge came to ask me to build an iWin and then, uh, look for someone else who did it and just went to help them. When I left Mindbridge, um, and then with Remedian, it passed. While I was on Mindbridge, there was already that accountant that came to me to help. But I was conflicted. I wanted to stay at Mindbridge. I couldn't go help anyone else. So, um, I decided after Mindbridge to play the mini solo GP game with seven LPs, um, all friends. And then one of the companies that came on my desk was Remittian, formerly called Taxpay. And the founder was an accountant who had the idea but didn't have a team, had a beta that was built by a friend of mine. And he said, sono, is this worth commercializing? Because that's costing us money. And accountants don't like things that cost money. And so basically, um, I started helping him and bought it off his hands. I, uh, took it out of the accounting firm. They had three accounting firms beta testing it. And then when I real. We did a market study, and my first reaction is that there must be someone who has done it before. My first reaction is, it's too simple. I mean, it's not simple to get all those licenses and all this work and build the agents to proper do compliance and all of that. But, um, my thought, my assumption was I can't do it with stripe. It's not worth my time. There's no mode. Someone will do it on it. And then when I dig under the layer and I sit down with him in his office and I understood all the rules because tax payments are not payments, they're compliance events. And that's why you cannot have a bipartite compliance level that's why any payment processors cannot do that that easily. It doesn't fit the market needs. When I uh, understood that and I realized that agentic AI is actually the only way you can do that at a reasonable cost at scale, and that the time now is ready for it, that's when I decided to dive in and get involved there.

Speaker B: Okay, can you just, let's bring this to life a little bit more tangibly. What does this mean? You have to have all these licenses, you have to have payments, there's a compliance layer, there's all this sort of statewide and compliance agents just paint the picture of like what it looked like before and what it looks like now.

Speaker A: So before um, there's 3,000 tax endpoints in North America. 3,000. It's not just the 50 states for personal and corporate and federal, there's also all different tax forms, there are all the different tax types and so there's 3,000 tax endpoints in North America. So without remission, if you're a high net worth bank in New York and you have uh, thousands of high net worth clients, guess what you need to do? You need to build a manual list of all the ACH accounts to be able to route this and the state or uh, tax authority will just change it in the middle of the tax season without you knowing. This year I can't share the name but it's one of the top 10 CPA firm calling us in distress because one of the top three wealth managers in the country decided to stop sending them pre filled checks. And also then they had thousands of returns to orchestrate for white glove service for their clients and they couldn't do it. So before remedian you had the best I can describe it is as if you're going on the Internet without Yahoo. Before Yahoo there was no repertoire and there was no Google. You know what I mean? There's no Google, there's no Yahoo. You need to know the websites. It's the same thing in the tax payment network today. You need to know where to send the money and that might change and you don't have any confirmation when it lands that you have sent the money so often. I'm hiring right now, uh, an employee from one of the leading tax authority in North America and he told me during the interview that 12% of tax payments are ah, misclassified by the internal systems and there's humans under tax receivership trying to quickly within two weeks reassign them properly before but the taxpayer still occurs two weeks of penalty. Although it's not their fault because there's no confirmation of payment, something as simple as that. So we do that, we avoid that. So even if the tax authority screws up, we can challenge them with a full deed confirmation.

Speaker B: I want to go back to a question we asked. Like, I think the obvious answer is, um, it's more complicated now. All these stories make sense, but this feels like a pretty big problem that still exists. Talk a little bit more about why this doesn't exist. It's 2026 right now. The concept of payments isn't a new thing. The concept of compliance is a new thing. Why is this so hard to solve? Why hasn't it been solved before?

Speaker A: Well, there was not enough pressure to solve it. Um, um, there was an executive order by this administration on September 1st of last year that says, no more checks for tax payments. So before that, there was a certain tolerance to checks. But checks are a great way to commit fraud. Massive fraud. Cartels launder money through tax authorities. Sometimes, like, it's crazy, they will send. This is a case I worked on where they will send a tax payment, but put a different ein for the return. So they will overpay and then they will get the refund back and that will launder the money. Huh. And they do that. And so they bought, uh, dozens of different companies and do that. So it's really something. That's why this administration has declared war on any money laundering activity, terrorists or cartel. The first thing they did is stop the checks and put a digital ledger at the federal level between the tax payment obligations. There's also another problem is tolerance. Right. There was a tolerance to late fees and penalties by the population. If I am, um, paying a late fee on penalty, well, it's not a big deal. But when the rate of penalties and the interest just jumped since COVID all of a sudden people are more sensitive to it.

Speaker B: Got it. So we have almost the kind of perfect storm of commercializable agents. An executive order that says, no more of this fraud. Um, and the market opportunity kind of comes together. It feels like one of those things where when we look back in history, it'll be like, oh, well, that makes sense why this massive company can be created, Created now. Um, the payments part and the executive order make sense. Um, I would say, you know, if I, if I look at people's websites, in almost every industry, accounting included, we see the word AI AI everywhere. Uh, uh, you know, it's like if you don't have AI, uh, that's what people are searching for. That's what People want. You talked a little bit about sort of this hybrid agentic system, uh, not AI bolt, like tell us a little bit more about what, what does that mean?

Speaker A: So when you do, um, payments, you have very little room for hallucinations. Edge cases like one, edge cases can cost us, you know, a lot of late fees and penalties if we take that liability. So we cannot afford the AI to make any mistakes. Right. So we have a system of systems that monitors champion challenger configuration in AI, which is also how we became so good at Mindbridge. Um, and with the champion challengers configuration, we can have AI uh competing with each other and you can have a human monitor all of it. And there's an escalation mechanism for it. So I would assume that about one in every 200pm, which is huge, there is an issue with the AI and that's why the humans gets involved right away. And um, and then sometimes it's also because of, you know, data sources. But AI systems are great at many things today. But that will, you know, even FSD on Teslas still requires human intervention for the edge cases. And those edge cases in payments are not forgivable. You have to have 100%. So that's why we build a hybrid mechanism. Both for the compliance agent we have, we uh, have a cross border agent on top of the payment agents. And soon we'll have a lending agent, an insurance agent and other agents. And I will, uh. Look, I've been AI for a long time. I know, it's amazing. I, uh, will need five times the stuff we have today to do what we do right without AI But I will still have a human in the loop and then a configuration of the AI systems where they compete with each other for the attention of the humans and for the accuracy.

Speaker B: So you said something that, um, I want to. It sounds very counter to a lot of the narrative that we hear, especially in the world of accounting, which is, um, I could have a huge team or I'm going to hire more people. Uh, isn't AI supposed to be doing the work? A team of 20 doing the work of 200. What does that mean you plan to hire more people? Uh, as an AI company, it means

Speaker A: that we will expand dramatically the usage and so the robustness of our DevOps, the ability to scale still requires more humans in the loop. We also are about. I don't know if I can announce it, but I mean, it's going to be announced soon. We're going to do a very large partnerships with one of the top three largest software Players in that space who has done the number one software, 65% of the market, which is them, for tax delivery. And so when you partner with someone like that, although they're not asking for it, we're going to create a clone of Remychen. We're going to create a fourth deployment infrastructure and given their volume is so big, we will have to staff up for supporting them. And, and you know what's very interesting with AI now is that people are starting to have a sense of data sovereignty. So even in the commercial agreements we sign, people make a sign that we don't have to train our AI system, we cannot train OI models on them or AI agents on them. So the only solution to that is to clone your infrastructure and give them full access to it and full authority over it, part of the partnership. So that's why for us also you have to keep in mind, okay, great, we did our first Pan American tax season but we cover 300 of the 3,000 tax endpoints in North America. And I'm going to go global. I already have requests for Vietnam, for Mexico from all the governments that struggle with the same issue. We are going to have a remittian, um, government edition. There's some countries that have very poor infrastructure for tax collection and we're going to have a product for them. So we will serve both end of the market, the taxpayers and the government.

Speaker B: So I want to pick up on another comment you just made. You just said you're going to partner with uh, you know, one of the, the top, the top players in the space. When you think about sort of the, the top players, there's kind of a lot of like I could build that. How, how hard could that be? I get if maybe you're sort of sort of like a non bay player. Got it. Tell us a little bit more. I was going to ask why, why pick a third party, uh, to do this, uh, give us a little more insight to like how that, how that sales cycle went.

Speaker A: Um, I have to say it was not a sales cycle, it was an information cycle. It was uh, sharing about what do we do, what do you do. Them trying to partner with other people and then their clients telling them remittian does it better, right? So um, I think if I was a first time founder, I would have been much more sheltered and protective over what I do by fear of the big ass clothing assets. The reality is that when you, you know that when you have a very small team of highly competent people that ah, are very obsessive seven days a Week on something. It's not a linear progression. It's an exponential production of outcome. Right. And large teams usually have committees, have budget, have a lot of things that make them. I mean, the reason why Tesla is Tesla and General Motors did not create a Tesla is a prime example of that. So one of the things I learned from one of my co founders at Mindbridge, when he joined the company, he had that quote on his, all his emails. Never underestimate the power of a small team. Convicts convinced that they can change the world. Highly motivated small team. It's a quote from a famous, uh, poet, I think. But that's exactly what happened at Remission is a group, the first 15 employees, we had some engineers that are just brilliant, that just do breakthrough and they just bang the head against a prime without any committee, without any boss telling them what to do or not to do. Exploring, testing at small scale. And then we tested at small scale, we did a larger scale. And that gave us an edge. Right.

Speaker B: You talked a little bit about Baker Tilly, Aprio, Krausoverman and other firms trusting you very early. Um, tell us a little bit about how they've sort of helped you either see the opportunity, expand the opportunity. Where does it go from kind of today going forward?

Speaker A: So it's extreme collaboration. We have two customer advisory board, and I consider them almost like my informal board. One of my clients was a big four when I was at Caseware and I had a very sad story. I witnessed where one of the partner I was working with had to take a flight in emergency to welcome the body of an accountant on a mission who died of overwalk at 33 years old. Stress. He died of stress at the desk. People don't understand what tax season is. And to me, if I receive an email from an accountant and depending on us, I know what it means behind, they have hundreds of clients. Stress for making the deadline. They're working seven days a week during the tax season. Like, I really respect that. So to me, each one of those clients, no matter how small it is, I take it really personally at heart when they have any issue with what we do. And so, uh, I don't even call it like a client relationship. I call it really like team collaborations. Like they're part of the team. I look after their experience with us very intimately. Um, I'm very active on those things. And so. But, um, they're also very conservative. The bar to meet their satisfaction is high. Right. They're very conservative buyers. And you have to accept it and respect it. So that's all I can tell you is that collaboration was also, I mean we have a Larry Award. So the passion patient zero, the accountant that had the idea, um, we have a Larry Award in the company for the person that represents the value that he represents, which is dedication to customer success, grinding in silence, being strategic and being ultimate client care. So we institutionalize that in the culture of a company with the Larry Award.

Speaker B: I love that. I love that. I have found in my various founder journeys. And now that there's a couple things. You go in as the founder of a company with sort of a view of the world and then there's these deep partnerships like you talk about, Baker Tilly Aprio, for example, where they have their view of the world and their view of the future and their constraints. What's something you went in that helped surprise them at, ah, what was possible? What's something that you went in that they helped sort of inform your roadmap that, that might have surprised or changed your roadmap along the way?

Speaker A: Um, I think what's very interesting is that, and I don't have to answer the question perfectly, but, um, they make a lot of assumptions of what their professional standards and the law can or can't do. And they're very conservative and we pushed a bit the needle. So when we went to see them, they thought that they were not allowed to do certain things by the law or by their standards. And I challenged that. And when we went even with the head council in their organizations and they said, no, this is actually okay by the law and you should be able to do that. Then when they came to us, they said, okay, but then we need to do this, this and this and that. So their comfort level is not just meeting the law, is making sure that the perception of the work they do by their clients eyes is always of the highest standard. And that was very interesting for us to see, which is not only is a certain convincing of the new practice, of the new method that needs to be used here, this conformity, there's certifications, there's a lot of things. But even when they go and implement it, they really want the least amount of stress, the least amount of hoops, um, that made us better. It took more time to launch the uh, product several times. But it also makes us better because when it's a click, um, and go, all the other firms come on it and it's just a standard approach to it. So they are very standard sops in accounting firms, the large ones at least, and the small ones, they cannot Survive, uh, without them. The way they do engagement letters, the way they engage with the client, they have to standardize everything otherwise it becomes chaos for them. And that really changed our approach to really fit into that.

Speaker B: Can you share a little bit more about this thing you were talking about with the general counsel where they're like, you're like no, you actually, you can do this.

Speaker A: Yeah. So taxpayer representative rights, um, is something that the taxpayers can delegate to the accountant to operate on their behalf, to access their records and to do certain things. And most CPA firms don't know how to use it or leverage it digitally properly because most of software vendors are not as stringent on certifications and security. And we are like triple audited anti money laundering Nacha and SoC2. Right. And so we are very comfortable with that conservatism and those steps involving it, um, so we're able to meet it. But some of that contest had no clue they were allowed to do that. Um, they never considered using this in that context. And that was one of the breakthrough that made us different.

Speaker B: So if we start to apply this towards sort of the broader market, what is it that some of those early firms, I think appreciators or saw that the rest of the market hasn't appreciated yet? What is it that Baker, Tilly and Aprio are ah, very large firms. What do you think they're seeing? If we sort of try and draw this lesson out to the, the other hundreds or thousands of accounting firms out there, what do you think they saw that the market hasn't appreciated yet about remediation?

Speaker A: Well first of all, um, the leadership and it's Baker to the RSG by the way. Um, so it always comes down to leadership. Like it's silly, it sounds cliche, but the leadership of Aprio has always been, that's their slogan I think, passionate for what's next. And they always have that thirst to find out how to push better client service, better use of technology. And so they really are very actively exploring possibilities. Right. But the other firms that don't necessarily have that four innovative culture like the firm like McIntyre and associate crossoverman and others, they know some of their clients struggle with it. They just know some. I mean do you think a 79 years old old lady widow is going to have the patience to learn how to register on seven different state portals to do her taxpayer on time? It stresses her to sit in front of a computer. Right. So there's those cases. But you think also the dentist, if you look at patient zero, Larry the accountant, his Wife Shana, um, Isabelle, a doctor. Okay. And he was preparing her tax return for her. And he was frustrated because sometimes she missed the deadlines. And he said, I don't understand. I gave you the spreadsheet. You know, and at home, imagine the discussion at dinner, the accountant says to the wife, why did you miss that? I gave you the spreadsheet at the beginning of the year. And then she keeps. And then he realized that, you know, liberal professions like realtors, doctors, lawyers, that customer facing are busy. They're not going to, you know, do once a month log in on the bank account. So one of our clients, one of our early clients of his firm, um, had a reminder on his iPhone every last Thursday of the month to go log in in three different banks to go pay the three different full co business co thing. And so there's these, there's those firms that know that they have clients and farmers. I mean we have a situation where it was not just the, uh, avail of remission, it was remittian plus Starlink because the farmers had poor Internet and the bank and the accountants will drive for hours to the countryside in the prairies to collect checks for their clients because the farmers are not with a laptop all day long in front of them.

Speaker B: What?

Speaker A: So, so you can imagine there's, there's a lot of situations like that people overlook. And those accounting firms had 20% of the clients that were those difficult clients that kept having penalty fees, not because they wanted to be outlaws, just because they're busy in their life, busy parents or busy small business entrepreneurs, which is the majority of their clients. And they just wanted a solution to alleviate them from that pain, which is a tax distraction, which is making the payments manually. And those accounting firms know all of the, when we go to see them, say, oh, they already have 7 clients, 10 clients, 20 clients per partner on top of mind, you know, and then corporations that have multi states, um, for example, right now we're talking to a business that has 900 locations across the US and they have 12 bookkeepers managing every quarterly estimates. Right. It's a lot of manual work that doesn't need to exist.

Speaker B: A little while ago you announced, uh, your API and funding almost simultaneously. I think after talking with you, it's easy to see the opportunity, it's easy to see why you are the founder to Beton. But like you said, also it's been quite a, uh, afterthought for the VC world to invest in this space. Tell us a little bit more about how that fundraising process went and then what is it that sort of you sort of did to help get them over the finish line of comfort? This is a conservative industry in a conservative market.

Speaker A: Um, I'm able to talk to the clients, I mean like for those who spend enough time on us to at least take phone call from one of our clients because you can say all day it's almost surreal to explain that problem and that it happens this way. But when they talk to some of the clients and they give them exact case studies and exact problem and then they realize the pattern that across all CPA firms it's about another. And then you drop one number 85 billion of late fees and penalties per year in North America. It's a tax payment network of $7 trillion in the US alone, $10 trillion in most countries we're going to go to right now and globally even bigger than that. And then also in clicks we are activating and digitizing a tax payment network that has never been done at that scale into trillions of dollars. And we're the first one to do it the right way. And what's exciting about our API product which now is the majority of our sales, there's that whole digitization and AI based tax engines now surging. April ah Tax and others I Win has a tax engine as well. And so the only answer to the lack of the shortage of our content is AI. And guess what? When those people like instead.com do an AI tax engine, the next thing that comes they want to because they're digitally native. Why would they not orchestrate the payment right away? And we're the first API that can do that.

Speaker B: So one of the other things that Silicon Valley is a proxy for the venture world in general likes is a wedge into the market that creates a bigger, broader vision. If Remedian succeeds in the way that you envision it, what do you think the tax payment experience looks like five years from now?

Speaker A: Seamless, painless. And there's no more deadlines. So let me explain that if we are intimately embedded in the financial life and the bank accounts and the tax data ah of the accountants and we have machine learning everywhere and AI to estimate their returns even before the deadline, why would I wait once a year to take a lump sum payment from them where I could spread the load over time with financial services. And that removes the pressure on accountants for the deadline too when they're short staffed. I want to kill the deadline of tax season. It's so stressful, it's stupid. There's many ways we can bear the load over time, create tax savings account up front, offer lending products, many things we have in the works right now to just make life easier for taxpayers. By the way, until remitian, I always had late fees and penalties because I have three or four entities and I'm a busy entrepreneur and have all the things going on. Right? So to me it's, it's, it's personal thing too that it just um, when I started using it, I was like, this year there was something interesting that happened and I posted on LinkedIn. All of a sudden I see it's May 1st. I'm like, oh shit, I forgot to pay my tax bill in Canada for that property, whatever. And then I, I, I emailed my accountant, I said, hey, you didn't see that? So it's taken care of. Uh, like, what do you mean you forgot you're on remittance? Like, I was like, I was like, oh, that's the first year that, you know, like I had anxiety of not paying a late fee. And I was like, oh yeah, that's the feeling of what I do every day. That felt very good. It was a bit like that bites the tail, but that's how it should be. People should not have that stress. You don't stress about paying your cell phone bill or whatever utility you have on your credit card. Why would you straight to pay the tax man. You should not have to stress paying the tax.

Speaker B: I love it, I love it. Uh, one of the other fun things to end on a little bit of a fun note, uh, so I see this a lot in the trend around and it's no different with you guys. You name your agents, tell us what are the names of your agents and how do you come up with those names and why?

Speaker A: So we have, my three favorite agents in the business is Magnus, Sterling and Lila. Um, Magnus is our compliance agent and because m, uh, uh, one of our engineering leads that is the most compliance oriented and head of security, his dog is a very strong lab called Magnus and he fetches very well. So Magnus Fetch. Um, Sterling was actually one of my co founders preference. I don't know why he picked that name but it's a finance agent, treasury agent and it sounds very sophisticated. Sterling picked up with it. I still don't know why I picked up that name, but I went with it. And Lila, it's because it's uh, the wife, one of the co founders, she's on top of the money. It's our Rev Ops agent, she always tracks things properly. Um, she's very on top of Things she never forgets the details so we call her Lila. So we just found personal references in the lives of the co founders and the team. Um, and uh, that's how we named it, like this. And then what we do is we took the pictures of all the people in those teams and created an AR avatar combination of all people in those teams as a face in Slack. Uh, so it has a name and a face.

Speaker B: Love it, Love it. All right, tell us who is your ideal customer right now and how should they get ahold of you?

Speaker A: Our ideal customers today are uh, CPA firms of um, more than 500 clients to more than 1,000 clients. Uh, they should get hold of us through our website remittian.com or through my LinkedIn Solana Angel. Um, on our website they can start a uh, free unlimited trial with a set of feed that is refundable and we can guide them very quickly. We also, the other ideal clients for us is technology companies that have remittances somewhere in their workflow. Instead of sending people to a state portal, instead of sending people a PDF with how much tax they owe, they should embed our API right away to make it seamless. And we have a partnership agreement with them where we do rev share as well. And then we have an emerging third category which is uh, banks, um, banks that service small business owners that have those portals where they can see you know, obligations on pay the taxes instead of just giving a drop down list of all the tax obligations. They can have Remittian UI embedded in the portal client portal for better client experience. So we have a few community savings banks and banks that we're talking to right now and uh, high net worth banks that orchestrate the payments from the um, you know, IAS and the Citizen and the wealth investments accounts. Um, so those are, those are the three people that should look at us today to make their clients life easier. Right. And of course sometimes we take direct clients. We have clients that change CPA firm but want to keep remittian. Right. And if the other CPA firms are adopting it, they pay us directly to keep the service going. So that's uh, a, for us that's a great, an emerging proof point that ultimately we're doing a good job of. The taxpayers are more loyal to Remedian um, than a CPA firm. That says a lot.

Speaker B: I love it, I love it. Well, congratulations. Thanks for sharing your story. Thanks for sharing your time uh, and wishing you the best of luck.

Speaker A: Thank you Sasha. Thank you for having me.

Speaker B: Thanks for listening to Tech Finance with Puzzle and Turpentine. If you got value from this episode, please share it with any anybody building or funding the future of technology. And please leave us a review wherever you're listening and tell me about what you want to hear more of. If you're looking for accounting software, you'll actually love to power your company. Check out Puzzle IO, uh, the modern accounting platform for the AI era. Tech finance is part of Turpentine, the awesome podcast network behind Econ 102 with Noah Smith, Turpentine VC Complex Systems with Patrick McKenzie, and so many more.

Related episodes across the Index

Other episodes covering the same guests and topics, from across The B2B Podcast Index.

  • Why AI Pilots Fail: How to Escape AI Pilot Purgatory and Scale Enterprise AI with Ronnie Kwesi ColemanUsing AI at Work · on AI agents92 / 100
  • More Agents Than Employees: How Zapier Disrupted Itself Before AI CouldTalking AI · on AI agents90 / 100
  • PMs Are Fintech: How Column is Rethinking Property Management BankingThe Profitable Property Management Podcast · on AI agents89 / 100
  • Inside an AI-powered personal dashboard | Matthew TiemannPossible · on AI agents83 / 100
  • 90% of AI prototypes never reach production (w/ Temporal's Samar Abbas) | AI BasicsThis Week in Startups · on AI agents81 / 100
  • From Billable Hours to Business Outcomes: How AI Is Rewriting the Rules of IT ConsultingEvolving the Enterprise · on AI agents78 / 100

More from "Tech Finance" with Sasha Orloff

All episodes →
  • E68: Jeff Grimes of Perplexity on the New AI-Native Analyst
  • E67: Navigating the Era of "AI Slop" in Accounting with Luke Frye
  • E66: Engineering the Modern Accounting Firm with Greg Capitolo of Attivo Partners
  • E65: a16z-backed Stuut CEO Tarek Alaruri on Why Legacy AR Software Failed and How AI Fixes It
  • E64: Numeral CEO Sam Ross: Why I Abandoned My Profitable Business to Raise $50M for Something "Boring"
Explore the best B2B Finance podcasts →
All "Tech Finance" with Sasha Orloff episodes →