AI Powered Tax Compliance and the Road to Crush 2026

Hey everyone, this is Alan Shimel here on Techstrong TV, and I'm about to interview Hugo Sarrazin. Hugo is fairly new CEO at a company called Avalara. This is going to be a series of interviews, actually, probably a three-parter at least, and it'll be leading up to something that we're going to be covering here.

It's actually an event down here in South Florida. It's called Crush 2026. It's the Crush 2026, it's a user conference for Avalara.

It's over at the Omni Resort, which is a beautiful, sparkling new resort down in Fort Lauderdale, and it's September 22nd to the 24th. So everything we're going to talk about here today and over the next couple of interviews I'm going to do with Hugo, really can be seen at Crush. So if you haven't made plans yet, and what Avalara does is something in your sweet spot with financials, accounting, and so forth, as well as AI, might be worth your time.

And the end of September is not a terrible time to come to Fort Lauderdale. It's always a good time to come to Fort Lauderdale. Anyway, let me introduce you to Hugo Sarrazin.

Hugo, welcome to Techstrong TV. It's great to have you on here. Thank you.

Thank you for having me. My pleasure. So Hugo, if you don't mind, I'm going to go to the glasses so I can read just a little bit.

As I mentioned, you're fairly new as the CEO here at Avalara. Give us a little bit of your journey. How'd you wind up here and when didd you come?

Okay. Well, I've been here a full 90 days. And before that, I was the CEO of a publicly traded company that merged with another one, and before that, another SaaS company where I was the chief product and technology officer, and before that, a quarter of a century, in the tech space, at McKinsey, where I led our Silicon Valley office and served just about everybody across the whole technology landscape.

So I'm a technologist at heart. I even have a PhD from Stanford in Robotics and AI from 30 years ago. So I feel like the moment is now.

It finally is happening. I couldn't be more excited. We were talking off camera.

I think for those of us who, as I said, we're always told that the Jetsons' flying car was five to 10 years out. We're finally getting to where it's not five to 10 years out anymore. So I think equally excited here.

Hugo, I feel like we wouldn't be doing justice to this if we didn't start off with telling our audience a little bit about Avalara. They may not be familiar. You're CEO all of 90 days.

How would you describe Avalara to the audience? Yeah. So we've been in business for 22 years.

We're the leader in sales and use tax here in the US, but also around the world. We have a sizable business in Europe and APAC. And what we help do is, when you have a transaction, we help calculate in real time the taxes that are due.

And in this country in particular, there's 1,200 jurisdiction that keeps changing and modifying and adapting, and it's incredibly, incredibly complex. So we help companies keep track of all these changes and then do that in real time with low latency. So all these e-commerce companies, when they need to kind of calculate tax and let the user know what it's going to cost to ship something directly to North Dakota or Louisiana or Colorado, we can help you.

But beyond that, because we know all the sales tax, we can help you file your return with the government. We can help you manage exemption because there's exemption certificate. When you do cross-border, we calculate cross-border tax.

Around the world, there's e-invoicing. So there's a lot. It's a fascinating and complex space, and one that is just continue to be more and more complex.

Not the last time as we've seen the tariffs are changing on a daily basis, and we're in the middle of all that. Absolutely. Fascinating for you.

I could tell you as the CEO founder of a company, when it would come tax time and we'd have to do the sales tax computations, and we don't even sell product per se. Well, everything's a product, but we don't sell physical product is what I'm trying to say. Yeah.

But the different jurisdictions and, especially when you're an internet business, right? Because a lot of them treat internet businesses a little funny. It is.

Complex doesn't begin to describe it, to tell you the truth. But at the same time, Hugo, like everything else today, AI is everywhere. This seems almost like a tailor-made kind of world for an AI helper.

Yeah. We can unpack it. They're different things.

And then at the scale we're at, I didn't even touch on that. We've got two million direct and indirect customers. Oh my God.

We're large, and we see the flow of transaction in and out. We see these tax return. We see what gets audited.

So there's a lot of data. So yeah, it lends itself to AI in a very, very nice way, but we need to be very thoughtful. Calculating taxes is not something that you leave to a probabilistic engine.

It needs to be the same answer every day, every time under the same condition. So we have at the heart a deterministic calc engine that does these calculation and will always do that correctly and in an auditable way, and a way that you can feel confident you're doing the right things. Now, around that calc engine, all the processes, all the things that are required, including monitoring the changes, you can use AI, and you can use AI to really improve the performance and transform a lot of business processes that have been put in place.

And that's what we're doing. Fantastic. I'm sitting listening to you and I'm thinking to myself, can you imagine the IRS calling you in and you figured this out wrong, and you said, "But I used the AI," and we log on to a Claude or a ChatGPT and say, "Hey, this is wrong.

" And the chatbot says, "Oh, my bad, Hugo. " We can't afford that when we're talking about things like tax calculations and so forth, right? You can't.

You absolutely can't. You need the audibility. You need to have the referenceable tax code.

And for all the great quality, and it's amazing what these public LLMs do. They're subject to data that's sitting over there, and they're scraping the internet and the Reddit and all the use group, and that's how they've done an amazing job at learning language and being able to solve complex problem. But I think when it comes to tax, you need very specific understanding.

So we not only do some of what I've just said, but we have people who are tax experts who have curated the data. You need to also understand how that data was used in filing, and then you need that feedback loop. And we've been doing this for 22 years, so we have data that is not available on the internet.

It has not been used to train LLM. It has been used to train our own version of the LLM, and that's why you can do this in a more accurate way. Secondly, you need to keep track of what you do.

You need to have this audit trail, and our platform does that for you. It keeps track of what tax code was used and what context and what precedent was used, and you can interrogate the platform and find out a whole lot about what jackets cost in New York State, which is, by the way, different than Boston State, above $150. And there are all these fancy dandy thing that you need to understand, and we can kind of give you the correct understanding and the audit trail that goes with it.

I love it. I'm sitting here thinking, the depth of it and the complexity of it can really kind of drag right at you. Hugo, we can't cover everything in 15 minutes, right?

Yeah. So we're going to talk about some things, and then we'll continue it in our next discussion. But AI, nevertheless, is fundamentally changing so many businesses, whether it's in law, medicine, accounting, technology, certainly.

You've outlined what some of the limitations are, but I feel like in the interest of fairness, we've got to say what a great thing, how much it helps as well. Oh, it is amazing. I want to spend my entire career getting ready for this moment, and I think it's a transformative technology.

It's the next leap forward that we can anticipate. We've done automation for years. But automation took existing processes and found ways to skip step or get step done automatically.

If you use AI correctly, you can reimagine work, right? You can kind of transform that work, and that, I think, is what's interesting. And to continue with the tax example, one way you use our product is in a ERP system, where you run your ledgers.

There's an out-of-the-box tax code, which is pretty basic. It doesn't mostly suffice, so people buy our product to augment that. And you configure that based on your understanding of what you need at that moment in time.

And that's the way it's been. But now if your business changes, you sell a new product, you sell to a new location, there are new tax rule, who updates that? Right?

Yeah. " We'll keep you in the loop, by the way. We'll tell you, "Hey, something happened in this state or in this jurisdiction, or the tariffs just changed.

Do you want to do the following sets of things? " Mm-hmm. It's kind of not them that.

You've had people manually go back and try to tweak and adapt and modify. Same thing, exemption certificate. An exemption certificate is if you do business with a certain entity, some of them have exemption in a state, in a jurisdiction because they're a not-for-profit, they're a church, or they're a wholesaler, whatever.

These certificates need to be kept current. They need to be read. They need to be uploaded.

All of that now you can do basic machine language. You don't need generative to kind of read a document, extract the character, and then manage that, and then keep that current. So there's a lot of reimagining of a lot of processes that can be put in place, and I think we're just seeing the beginning.

But the core thing I just want to highlight is you need to be ready to change your operating system. You need to reimagine the work. You need to invest in your people.

You need to reskill them to operate this way, because if you don't understand AI, if you're not able to interrogate, if you take the answer coming out of your AI as always the right answer, you're not being critical enough. And I think there's a lot of really important new reskilling we need to do of the workforce to get people to be good at taking the output, asking the right question, and then refining. And I think organizations have a huge opportunity to do that.

I agree with you. You know what? I told someone this this morning when I was on a video.

This is a moment in time. You can't take a snapshot right now and say, "This is what it's going to be like forever," even six months from now. This is a moment in time along this evolutionary path that we're all walking on around AI.

And what makes sense now may not make sense in just a few months. And what we're doing now may not make sense in just a few months. But Hugo, we're going to talk more about this in future ones.

But we got a minute or two and a half minutes here. I want you to come back to Crush. Yeah.

Right? It does kick off, as I said, September 22nd, down here in Fort Lauderdale, especially for those of you watching this in the South Florida or anywhere from Tampa on down. This is worth the drive, right- Oh, yeah ...

to come to and check it out. Can you give us-- Let's not get in trouble, but give us a preview of what people can expect from the main stage from you and some of the key speakers. Yeah, listen, it's our annual user conference.

We got a version in North America on the 22nd. We have a version in London on October 8th. We bring our users, we bring our partners, and we do a few things.

One, we give them a bit of a state of the industry. Where are we today, what to expect, and what gets us excited, what gives us pause for concern. So there's going to be some keynotes to do that.

We have a product that's already very agentic, but we're making a number of announcements. So you don't want to miss out on a lot of them. I'm not going to spoil the fun.

But it's pretty amazing. The team is busy in the engine room tweaking and adapting and refining all these product capabilities. So you want to get a glimpse of where the future is going.

We're going to be sharing a lot of that. And we're going to give you a sense of what the product is capable of doing now, what it will be capable of doing at six months, and as you said, it keeps changing. So we're preparing for more change and letting people know that.

There's going to be tracks for different sub-product, depending on your needs. There'll be some customers who will share their experience, some of the beta users and some others. We'll have partners who will also share how they're helping companies adopt agentic capabilities and transform their work product.

So it's kind of a nice moment to step out of the day-to-day grind, go and learn about what's happening and what's coming, network with other peers of yours that have the same needs or experiencing some of the same challenges, and just get better and get ready because-- I say that internally with our own team. The future is for people who are good at learning and adapting and changing, and everybody has to have that mindset. So this is an investment in yourself.

It's an investment in the company. It's preparing you to continue to be relevant as we continue to change. I love it.

Hugo, thanks for coming on here. I look forward to continuing the discussion between now and the 22nd, and then continuing it in person with you- Absolutely ... as I said, at the Omni Hotel down there in Fort Lauderdale.

Sounds exciting. As we were talking, what a great time. An amazing time.

Who would've thought, right, that we'd be where we are at this point in time, being able to just kind of watch this and partake in it unfolding, so fantastic stuff. Good luck and success in your role here as CEO at Avalara, and we'll continue this in a very short time. Excellent.

Thank you. Thank you. Hugo Sarrazin, CEO of Avalara.

Their Crush 2026 event, as we said, down here in Lauderdale, September 22nd to the 24th. If you are involved in the financial industry, especially anything to do with sales tax calculations and so forth, this is a can't-miss one, so check it out. We're going to take a break here on Techstrong TV.

We'll be back with more soon.