Why AI Demands a New Approach to Financial Systems of Record

[00:00:00] PREVIEW: So the macro right now of what's going on is sort of AI. You do need clean data and a clean source system of record data to get anything done because otherwise it's a little bit of a garbage in, garbage out type of thing. These systems of records, they're literally 20-30 years old in some cases, the modern version.

So on a data side, you have the issue with the systems today that they ironically have stopped becoming the source system of truth in many cases. So you have all these upstream systems that hold a lot of that data and then pump down, like quote-unquote, dumb journal entries into a dumb general ledger. Again, air quotes here for the folks only listening. 

If you get to a platform that's well-integrated, well-systematized, that has a lot of core, valuable data, then all of a sudden you can enable a ton of interesting workflows to automate a bunch of things. So yes, it is a source system of record, but that's kind of the charm of the beauty, especially in this era of AI.

You can actually build a net new product that's highly differentiated, faster to implement, faster to do the workflows, and better reporting that are strong enough for people to actually move off their old system.

[00:01:00] INTRO: Welcome to Unpack Pricing, the show that deconstructs the dark arts of SaaS pricing and packaging. I'm your host, Scott Woody, co-founder and CEO of Metronome. In each episode, you'll learn how the best leaders in tech are turning pricing into a key driver for revenue growth. Let's dive in.

[00:01:24] Scott: Welcome to Unpack Pricing. I'm here with our special guest, Nic Kopp, CEO and founder of Rillet. Nic, I would love to just start by maybe having you introduce yourself. Would love to just say like, who are you, what is Rillet and why are you like, so crazy? Like, what is wrong with your brain such that you decided to start an ERP in 2025 or in the mid-twenties? 

[00:01:43] Nicolas: Yeah. Sounds good. Yeah. Not, not everybody jumps out of bed in the morning, wanting to build an ERP system from the ground up. So here we are. Yeah, excited to be here Scott, thanks for having us. So, I'm Nicolas Kopp, CEO and founder of Rillet.

We are building an AI-native ERP from the ground up. So that's every single building block, every single debit and credit, full general ledgers, subledgers, all built from the ground up in an AI-first way. I'm very excited to share more there. The company is three-and-a-half years old and it took us a little bit to build the foundation here.

And came outta stealth roughly one, one-and-a-half years ago, backed by Sequoia, Andreessen and Iconic. And yeah, very happy to have many customers already. 

[00:02:23] Scott: Awesome. Okay, so I think when you talk to entrepreneurs, you hear lots of advice about how to think about starting a company, and I think, the ERP to me at least, kind of occupies this very interesting space where it is held up as a paragon of, like, a bulletproof system. Like, a system that you strive to one day hopefully have the GRR that approaches an ERP and it's like indisplaceable. And so it's this like, theoretical construct, at least in my head. It's like, the most bulletproof part of the software stack. And yet, it is now [00:03:00] 2025 and you're building something that is going to attack the kind of core foundation of every business on Earth, which is the ERP system.

[00:03:06] Talk to me a little bit about how you're thinking about that problem and in particular, why now is the right time to kind of turn over, like, have a new generation of ERP? Mm-hmm. 

[00:03:18] Nicolas: Yeah, great question. So, the why now is actually very intricately glued together with sort your question around like, okay, these are very core systems of record that hold a lot of data.

So, the macro right now of what's going on is sort of AI. That part is, maybe, obvious. A lot of our companies these days are rooted in that ground truth. So, there's a lot of workflows and automations you can build today with AI. Even in accounting, we can talk about that as well.

But, you do need clean data and a clean source system of record data to get anything done. Because, otherwise it's a little bit of a... what we're gonna do: the kids-friendly version, garbage-in-garbage-out, type of thing, otherwise. So you have, on the one side, these systems of records, they're literally 20-30 years old. In some cases, the modern version. So the cloud era has dominated sort of the last change there.

And for us, having a system that's well-organized, well-structured is sort of the ingredient for AI and the AI-enabled workflows. So on a data side, you have the issue with the systems today that they ironically have stopped becoming the source system of truth in many cases.

Mm-hmm. So you have all these upstream systems that hold a lot of that data and then pump down, like quote-unquote, dumb journal entries into a dumb general ledger. Again, air quotes here for the folks only listening. 

In this day and age, if you get to a platform that's well-integrated, well-systematized, that has a lot of core valuable data, then all of a sudden you can enable a ton of interesting workflows, um, deterministic workflows even, and AI-enabled workflows, to automate a bunch of things. 

So, yes, it is a source system of record. But that's kind of the charm of the beauty, especially in this era of AI. You can actually build a net new product that's highly differentiated, faster to implement, faster to do the workflows and better reporting that are strong enough for people to actually move off their old systems.

[00:05:02] Scott: Awesome. So, let me repeat that back 'cause I think it's like, subtle but seems important which is that basically in order to build AI-enabled workflows, you need a really clean source of truth data set. And the challenge with the old school ERPs is that basically for, at least your market, they don't have the source of truth.

They're so hard to use or they're so like... basically they have like a subset of the data that you need. Maybe they're like, useful for the general ledger, but they're kind of getting a fraction of the data that actually needs to happen in order to execute these workflows. Is that first part right? And then I have a second part.

[00:05:35] Nicolas: Yes. 

[00:05:35] Scott: Okay, great. So in a sense it's like, by building a really great product, you are now able to ingest high quality data, which will then enable the AI work-based workflows. Okay. That seems really... 

[00:05:48] Nicolas: Just to maybe put a story on this for folks. I mean, if you're familiar with ERP, then hopefully, what we just shared feels quite intuitive.

If you're not just for context like these ERP systems are again, they have been put in 10-20 years ago, as in some cases, there are layers and layers of like middleware consulting spend that build custom connections as the business model then moves on , you're in a situation where all of a sudden, like, all your old custom coding connections don't work anymore. You re-implement, putting new consultants on there.

But it's kind of a never-ending story of spend, and dollars and time where people just keep on patching the data pipes into these systems when if you rethink this thing from the ground up, things have moved on to a place where. A lot of modern tech companies are today on the Salesforce or a HubSpot or a Stripe or a Metronome. You have areas there where you have very dedicated, clean, upstream system of records that are able to pump data and buy their clean APIs into an ERP.

Yeah. And so you kind of leapfrog that sort of generation of middleware a little bit. 

[00:06:51] Scott: Exactly. Well yeah, so I think this is like, subtle but super important. When we started Metronome, we obviously thought a lot about ERP just because it's like a... it's like an end point for the data that comes out of Metronome.

And actually our first investor is the former CTO of SAP, Quentin. And, he told me the story about what is like... basically, I asked him like, what is ERP? I don't understand this software. It's like, this like, universal totalizing software system.

We worked together at Dropbox and I was like, did Dropbox have an ERP? 'Cause the way he would describe it, it's like the central source of truth. And he's like, well, we had... had a system that we called an ERP, but it did not function like an ERP in like, the old school way.

And then he described this story about Toyota, where basically Toyota is like a big SAP shop. And if SAP goes down, Toyota's badges don't work. Employees don't get paid. Cars don't ship. You can't buy anything. You can't even get into the building. And he's like, that's an ERP, system that like, coordinates everything.

And he is like, and at Dropbox, like just frankly we didn't have a system that was that central, right? Like, we had a general ledger of course. We procured NetSuite and all that stuff. And so when I heard that, I'm like, wow, this ERP thing, it's like, it is crazy. It's like this giant ambition.

How do you view where ERP is going? Is it going to be that thing that like eventually everyone is dependent on in the same way that Toyota is? Or is it more of a different scope than that?

And in particular, maybe use this as a chance to talk a little bit about which markets are you like particularly focused on right now. 

[00:08:26] Nicolas: Great question. And, people sometimes ask me as well, like, why do you call yourself an ERP? Because it's such an emotionally-charged term. So either people have a little bit of your reaction, Scott, or like, what is this thing?

But the only usual emotion you have associated with it is probably more like, negative than positive. And, for some people it's highly emotionally-charged because they had such bad experiences. So, we chose to build a subset of an ERP today. You can't, like, we would not legitimately be able to support Toyota today as a customer.

Not yet we will eventually get to this type of scale. But, so for starters, we are a general ledger at its core. So our direct competitor here is the NetSuite Oracle product that people, I'm sure, are very familiar with. General ledger, there's some revenue components to it that were revenue recognition, and all the revenue accounting workflows.

There's multi-entity multi-currency consolidation workflows embedded in there. As well as strong reporting capabilities overall, all tied together with native integrations, AI workflows. I'm gonna skip it at that abstraction level, but like, just for the folks out there that are familiar with the space, that's sort of the product suite that is embedded in ERP today, um, of Rellit.

And then from there, there's a lot of other mass and product suites and verticals out there. And some we will decide to eventually expand into and do, and others we won't. And examples of what we will quote unquote never do is, a great example is HR.

So in traditional ERP system, you will actually have HR payroll natively embedded in the platform of an ERP. We have phenomenal systems out there that currently do that as a standalone application. Same with CRMs. The NetSuite actually has a CRM still to date, actually, as part of the platform.

Again, realistically, all everybody's using Salesforce and HubSpot these days where some of these chunks of product, I think will not be reinvented in ERP and in other areas and surface areas of products that we're building around agents that can help with audits, for example, and going through audits. Or agents that can help with board reporting, investor communication.

These are products that all of a sudden an ERP doesn't have today, and so we will reframe and reshape the definition of an ERP, change the surface area a little bit, and in general, be much more communicative to the upstream and downstream ecosystems of data in and out , which is, I think, parts of the Achilles' heel of the current systems today.

[00:10:34] Scott: You're saying actually like, let's stay focused on our core and we'll layer on from there. Okay. I would love to then talk about that core focus in particular. I think, one of the interesting, you know, just truths about like this generation of AI. Obviously there's lots of different things that could be called AI, but like, you know, the meme about AI right now is gen AI and therefore it's stochastic and therefore it's not like non-deterministic.

And a non-deterministic in finance is like a very, uh, fraught combination. So, talk a little bit about how you see.

AI, working in terms of like, in the finance stack and then also, let's assume that you have this super clean data. Like, what are the areas where you're seeing the highest leverage in the finance suite for some of this gen AI stuff? 

[00:11:25] Nicolas: Yeah. So gen AI is famously bad also with numbers, accounting.

Yeah. And finance is all about numbers. So, to sort of get a bit of context and take a step back on, on your question on like, how do you apply AI in this sort of clean data world? We are an AI native system. We still have a lot of our accounting, core accounting workflows are highly deterministically executed.

And I'm curious actually how you think about this, , in the context of Metronome Scott, maybe as well. But like... if you, for example, close your books... so you have a accounting team of five to 10 people of your 200-300 million revenue run rate company. If you're in Rellit, you can support even a billion at that, this sort of staffing level.

And these folks will go through hundreds of individual processes to close their books, get the financials ready, reconcile financial information, and so forth. All these processes, or most of these processes can be broken down through three separate buckets.

There's a bucket of data ingestions. We just talked about that along the first part of this podcast a little bit. There's a middle part of calculations, and then it is the output side once the calculations are done where information gets analyzed and structured and communicated.

On the bookends on the way in and on the way out, probabilistic LLMs are very powerful for some of the workflows. The middle piece, highly deterministic. We personally don't use AI, and probably also other software builders not recommend using AI in this state of form. And so to be very tactical in a use case for example, is LLMs are a really powerful at reading through unstructured data and PDFs.

So if you want to amortize a prepaid, which is a process for the non-accountants out there of making sure software, if you have Metronome... for example, I'm a customer of Metronome's, I'm paying Metronome or a certain amount of money over a year, we wanna amortize these expenses over the course of that year. And to basically follow accounting rules.

So on the way in, you can read a PDF, you can get the service start and end dates of my Metronome contract here. When does the service start? When does the service end? We pick that data out in the PDF, we then feed it into the deterministic engine to calculate the amortization rundown of that expense that is highly deterministic. That's the middle piece. And then on the way out, again, once you have hundreds of software contracts, you can use LLMs to analyze the data, compare and contrast what's changed, summarize the information. So that is sort of maybe a little bit of a long-winded description though.

But I think very important for people to realize that in accounting and finance, it's really the ingestion and the output side where LLMs can be beautifully applied. The middle pieces, we struggle or not even struggle. I think, we don't think it makes 

[00:13:53] Scott: sense. Yeah. 

[00:13:54] Nicolas: Um, 

[00:13:54] Scott: yeah. I mean, I think roughly we take a similar approach, right, which is to say the core math of computing and invoice or doing, it's like, gen AI doesn't even have... there's no value there. Even if you could somehow shoehorn it in. Like, you know, we have a P100 guarantee on accuracy. So it's one of these things where that just doesn't quite make sense. But on the parts where a human needs to digitize a contract, or kind of interpret this contract and then translate it into a contractual tructure that's being run by a deterministic engine. Yeah. Like, a human has to do that. There's some room for error, right? And so in a way, like the LLM is just like a supercharged, faster human and it's like, yeah, they can make API calls against this stuff, and they can configure this stuff. And as long as you have workflows that allow, like error checking by a person, it's like more or less just better than using a human for doing this stuff that otherwise a human would do. And, then as long as you don't think that it has the deterministic property, you kind of look at it like the output of any normal person and you have the right error checking.

It's like actually a supercharge version of that person in terms of its efficacy or speed or whatever you wanna measure. It's like more or less like build the core, make it bulletproof, make it exactly accurate, and then define like good APIs that agents in the future will be able to use.

Like, that's a like very classic, I think, enterprise software way of approaching it. And it kind of gets the benefits of automation without the like, without trying to pretend like you're doing some non-deterministic math or something like that. That's not how we do it. Ok, that sounds very aligned.

How do you think about where AI is taking the finance function?

I think specifically what I'm curious about is, you know, I've heard some companies kind of like in software engineering, it's super obvious how AI is augmenting humans in like who writes software? How is AI or how do you think AI and agents are gonna start to affect the different parts of the finance function?

And you know, feel free to break down the different like finance roles, but like roughly where do you project that AI is gonna have the biggest impact inside of finance in the next like one to two years? 

[00:15:58] Nicolas: Yeah. So, we are sort of core focused on accounting even so I can scope it down to that area. Happy to talk about FP&A and the other finance functions too. But I think, let me talk about with what I see on the field, basically every day. So AI, so there's like layers of accountants workflow at month end. There is a lot of distractions also on that person. Just for context, like people have this worldview of an accountant sitting somewhere in a basement, going through files like that is totally not true. And folks and individuals, our customers, they have to really communicate well. They're like, these are cross-functional piece of work. Obviously there's individual pieces that happen in a calculate like theistic manner, calculations on a spreadsheet or a desk somewhere.

But there's a lot of communication involved, and especially in fast moving companies, you, need to communicate.

So, the first layer where I see AI having impact today already is on that communication layer. And this is very simple, like, all of us out there, like AI can help write emails and letters or respond to certain things, summarize certain things.

So that's a great way to extract some of the noise. The other area where we see more accounting related is. People get pinged a lot internally on question on what are expense policies? Can you help me explain my equity plan, and what happens if the company does this or that?

Or, has this customer paid us or not? There is some sort of like just first-level abstraction that I think AI is having impact today. I would encourage, hopefully, folks that you have beautiful case studies and success stories there.

Then the next layer down, in terms of AI adoption, in the next one year is where AI can really start executing end-to-end workflows. And that's what I'm personally most excited about in terms of what can happen. So very specifically, I can pick out an agentic workflow that we have in our application and software where accountants often at month-end reconcile and code cash transactions.

So, if you have hundreds and thousands of bank transactions coming in, you need to map these against certain other objects, invoices, expense objects, , other areas, bank transfers. And a lot of this can be automated today already with upstream systems. But there are always like a hundred, 200 that are left over that you kind of need to go through manually.

And that's where you can today with our software, but in general, I think this will be generally applicable shortly, run like research analysis on these transactions, can then look at your own chart of accounts and department structures and make suggestions on how these things should be saved in.

So all of a sudden you have a two-step process where you can literally mimic human's workflow that can run through with that example. And, maybe cap it here in terms of details, but like there are hundreds of these processes where it can start running, end-to-end workflows.

That's sort of the next, I think layer that's to come. And then the last thing, on top of that that's already happening as well is once these workflows run through, you can have agents that check the work on the backend. And that's sort of the third layer of the work of like comparing periods.

AI is really powerful. They're already on flux analysis for people familiar with that concept where it compares today's month versus last month and what's different and why basically, and so you have a three-layered cake, so you have the noise. I think we can all use different toolings there to like block that out a little bit to help answer some of these questions.

Then you have the workflows. I think that's where a lot of leverage will come in. And then the final step is to review workflow. I think that is already successfully being applied across as well. 

[00:19:10] Scott: Very cool. So, I'm actually really interested because you work with, you have some really impressive logos on your website.

You work with a lot of really great folks. You also work with smaller companies too. And, I think one of the things that I've personally observed is that, it's kind of obvious when you say it, but it's like very interesting is that if you just like spend a lot of time with smaller companies, you kind of get a peak into like what the future's gonna look like.

It's like they work differently, like if you work with Cursor or whatever, you just see that the way that team operates, it's just... it's just different. And, not to say it's good or bad, it's just different. And so I'm like really curious like you must have some examples of like finance teams that are probably working on, you know, maybe it's like not even really basically possible with the technology. They're kind of like, kluging it together, but you're like getting a glimpse into the future like, maybe for some of our folks who are larger companies, like, is there anything interesting you're seeing there?

Like, glimmers of here's actually a where the world is going, it's just that this like one finance person in this like ten-person startup is doing it just like a couple years early. Anything interesting that you're seeing there? 

[00:20:11] Nicolas: Yeah, I think what, and you mentioned Cursor with they're famous for the velocity and speed of what they're doing.

Mm-hmm. Um, uh, Metronome, I'm sure the same. We see some of these AI companies just being hyper-optimized on speed. Everything's optimized on speed, and so, you can't do that once you're an organization with 3,000 or 4,000 people. So, I'm not suggesting anyone should, but like, just as an inspiration.

We just had a customer that actually messaged us yesterday, smaller company, maybe a hundred employees. They, with two-people finance team, closed their books on the first business day at 1:35 PM. So literally in four to five hours of a working day , and so I think you will just have people obsessing over and being proud of how fast they work.

[00:20:56] The team itself, they're highly sophisticated individuals. They will upstream have done the data wiring, dot data mapping, like they, they've prepared, they've done this in preparation. During the implementation, they map of their schemas. But once you're there, once the thing is optimized, you're now at literally 1:35 PM like right after your whatever, when you have lunch, somewhere there, where you're done. And so that's sort of, I think, what I see the most successful companies do these days that comes with different hiring methodologies on who you maybe hire. This comes with maybe different bar for process perfection and messiness and like, you know, a lot of messages going back and forth and culturally how people work. 

[00:21:31] Scott: Actually this is, I hadn't planned to talk about this, but this's super interesting because we see the same thing, but it's in a different way. So, historically pricing changes at, so like, the reason I started Metronome was 'cause like at Dropbox, a pricing change would take like six months or something.

Just like technically to get it live and it was just like so painful. And, we would hit stat sig in like two days, but it would take six months to run the experiment. Um, okay. Terrible. Okay. Now our AI native companies, they'll do a pricing change. Like they'll announce a product at like 11 in the morning and they'll start pricing two hours later.

Now, they don't even enter Metronome until like two hours after the product is launched, which literally in prior generations it would take six months, months, quarters... It is horrible. It's like always the long haul. And so you're like, 'Okay, cool. Wow. The software must be amazing'. And yes, okay, the software's better.

But the thing that is like that you said, which I think is right, is that. The company in some sense is optimizing for how quickly and automatically and efficiently they can do the thing, which in this case is a pricing change. Or in your case it's like, uh, closing the books and it's a mindset shift. 

I would love for you to talk a little bit about like, what does this imply about who these people are hiring or the skills or the like attitudes or whatever, like, go a little bit deeper on that because I actually think speed, it's like a proxy metric. Like speed by itself doesn't do anything, but it is correlated with these companies that are growing really fast. So talk to me a little bit about like what does that, like in your head, [00:23:00] like what does that changing about what it means to hire for finance and for these accounting workflows and stuff like that. Like what's different? What what are you giving, what are you d-weighting and what are you up weighting? 

[00:23:08] Nicolas: Yeah. D-weighting, let's start there. That's maybe the, most obvious one is just like historical knowledge and expertise. So if you've done something for 10 to 20 years, when my parents, uh, grew up.

That was like the best thing basically. I think with Scott, you and I, probably this generation where it had some value and like, sometimes people were overdoing it a little bit and already had a bit of a negative signal. I think by now it's potentially tilted a little bit the other way where you, like can this person adjust to a complete paradigm shift on how to do work?

And so I think that's one area maybe to introspect and it obviously shows up in the CV, but it's more of a mindset again of like how adaptable people are. So, I think that's probably something that gets deprioritized. We see that in our own hiring. We under index on that sort of expertise quote unquote. You need to have a baseline, otherwise you can't do build good accounting products, right? Just to be very clear. But sort of like the very long tenured curate track gets maybe under indexed on. And the number one thing that we hire for internally is just resourcefulness. And curiosity. And these are sort of intricately, English as a second language, so I forget the term, but like they're intricately put together.

If you're highly curious, that helps you be relentlessly resourceful. How to test that, that's really, really hard. If you have a tip, maybe Scott, you'll know better. But it's a matter of examples on how they sort of figured things out in the past. Often these folks skew also a little bit younger, , potentially in their careers, , have to prove themselves more early on.

But yeah. So , relentless resourcefulness is sort of the number one hiring criteria currently not for us. 

[00:24:36] Scott: Yeah. I mean, like I still interview everyone and my question is just tell me about the, like most ambitious, the hardest thing you've ever done.

You know, I actually, sometimes I'll frame it as like, tell me about something that you did that was impossible. And they'll like go in deep and the people who crush that interview or the people who are just like, yeah, it's like I, you know, they found this weird exploit and they just didn't take the conventional wisdom too seriously.

And they went and did it anyway and figured it out. Like, to me, that is the... that is the AI age. It's like every preconception is being turned over. And so you just need someone who's gonna update their learning in real time to, the thing and know that in two months unfortunately, it's gonna be different.

And so I say like, that's the exciting part, you know, at least for me as an entrepreneur, that's the part that makes me excited. Okay.

But ERP is also the place where the slowest change is supposedly happening. So, talk a little bit about what, um, what you're seeing in the market, right? Again, I think it's like super, it's super interesting that like, if you were to define one piece of software, it's like bedrock for a company. You know, if you're a tech company, it's database. Okay. But if it's like, if you're not a tech company, it's the ERP and in as metric and software, right?

So what is actually happening in the market? Like, you know, are larger companies actually starting to question this or is it more like, the new companies, they're just gonna grow up on Rillet, they're never gonna not a modern ERP? Like, it even matter to you if you like, have people move from SAP, Oracle NetSuite to Rillet? Or is it like, actually I'm just gonna service the future companies? 

[00:26:08] Nicolas: Yeah. We have a large contingent of very fast-growing companies on our platform. That was also at the way we got started, as you can imagine. So nobody wants to buy an ERP from startup. It's just not a thing in the early days, especially.

Then people ask you about reference customers. I'm sure you had the same thing in the early days. They're like, there are none. Like we have zero customer, like Exactly. Uh, my, my parents really believe in what we do. Um, And so, that's a starting point you through creative ways and again, some customer desperation.

Inspiring vision. People take a bet on the team, you get to your first set of customers. So we still have a, yeah, large contingent of these quickly growing companies. We had, for example, went through a phenomenal growth journey with Windsurf as they exploded absolutely, they managed to scale to a hundred million beyond with like literally ahead of finance and controller and our system.

We just on the board of Mercor who's on their path to a billion here shortly. So you have these tech companies that are just like, okay, this is the new way. I don't have time, resources, the patience, frankly, to input an old ERP system. Then in terms of the migrations off the NetSuites and the bigger systems.

So, 30% of our customer base comes from there, uh, today as well. That took a little longer. I'm very proud though that we are now at this stage where we can absorb these customers, provide them a better experience than these legacy folks. But we'll, I think, always take or lag in terms of adoption, just, as the insertion points are less well-defined as, for example, with first ICP it's very clear once you start outgrowing QuickBooks and these systems, you need something new that's very well defined. For the legacy folks, it'll take a little longer. There's such a huge amount of surface areas, but we have interest from like, I can't mention these.

These names here public on the podcast. Some of them also have not signed up yet, but they're very large populistic companies that take a very deep interest in sort of what we're doing. And, to an extent some of them are too big where I'm like, guys, you're not, we are not ready for you guys.

I appreciate the interest. But there is a cohort of companies there, and I think it's just a matter of time. 

[00:28:00] Scott: Yeah. Like, so when you're thinking about it, so like we have a similar dynamic and for these core central source of truth systems, how does the finance team reason about it? Is it like a rip and replace at that scale, or is it like a thing that lives alongside, and it just services different, like a certain subset of their product suite? Like generally like when you're working with these larger companies, how are you approaching it?

And the reason I'm asking, so I wanna talk a little bit more about modern enterprise sales, but like, I'm just curious, like just factually on the ground, how you approach it. And then I would love to talk a little bit about your go-to-market motion. 

[00:28:33] Nicolas: Yeah, it's all about pilots. That's the TLDR. So at the end of the day, working very closely over an extended period of time with customers day-to-day like some of these folks in person, , as well where you fly out, like go actually, , spend some time together.

So it's, very much a, human selling motion. Yes, you're selling an AI product, , the hot like or cool software. But at the end of the day, it's very much a human human-based approach. And that's how we. Like, work together, build together. , And that creates fruitful partnerships basically.

[00:29:03] Scott: Awesome. And how, I guess, , maybe talk a little bit about building a modern.

I would say like SAP, Oracle NetSuite, you know, their products mattered, but their sales teams were legendary, is like what I would say. Like their go-to-market teams were truly incredible and they're all the leaders of the valley sales, like they all worked at Oracle. At one point, like all of that.

So talk to me a little bit about how you're thinking about that. Like, you know, you're building a product, a very product-centric company. I can tell just based on the way you're talking about it, but how are you building out your go-to market function, and how are you like pairing those two?

[00:29:42] Nicolas: Yeah. So. You detected? Well, sort of into, I think as a, team, we're very much accounting, engineering products. So like pride in the craft of what we're building. , You do, to win this market, you need to be hyper competitive on the go to market side because to your point, these are literally some of the most sophisticated sales organizations in the world that we're going against.

So, in terms of how to do that best, number one, it all starts with a sort of strong product. I think that part of like the product, orientedness, mindness, there's Ramp is a great company, for example, in the expense management space. They just have, they have a phenomenal go to market team, but like they have a phenomenal product that underpins everything else and there are, I wouldn't call it low-hanging fruit because some of these folks that Oracle's SAPs and so forth, they have built robust systems. But there are some... the differentiation on the product side is very clear when you go through the demo and everything else. So I think that's one big area that underpins everything that then leads to happy customers.

We have an NPS of 71. Mm-hmm. NetSuite has an NPS of -3 Okay. And so that in itself, I think, gives a lot of organic love and ecosystem, a lot of chatter. CFOs are very referentials, controllers are very referentials when they buy. And so from that perspective, I think that's sort of the... the underpinning is a good product, happy customer that fuel into an ecosystem and environment that is frankly, notoriously low NPS and bad products. From there, it gets accompanied with sort of a little bit of a brand halo I think that we did a pretty decent job at, of just like amplifying our customers, amplifying all the good work that we're doing. We have a clear narrative on LinkedIn that is a very powerful medium for us to talk about our story.

We're not even, we just share all the wings we have, , and how we're building the company that amplifies, , that, that then gets paired within. Excellent effing in sales team, they're just like hyper custom, hyper consultative, but also holds customers and prospects accountable for like going through the sales cycle in a sort of like efficient manner, let's put it this way.

Yeah. And sort of that's the final spear then of the crew on the go to market side, but it's all customers and brand basically. 

[00:31:49] Scott: Yeah, I actually love it. So, talk to me a little bit about how you discovered the right way or how you're discovering, because I definitely, you know, I'm sure there's a lot left to learn, but like how you're discovering the right way to position and sell your product in this market? Like what are the things that you are doing or your team is doing?

Like how are you creating those learning loops? Because you're essentially rediscovering the way to sell ERP. It's like there was a playbook, it was written 20 years ago when NetSuite came up. Now you're rewriting it. So like what are you doing just actively to kind of promote that learning and like how are you like actually getting better?

Or how are you and your team getting better over time? Like, what is the mechanism of doing that? 

[00:32:27] Nicolas: Yeah. We do, and this sounds maybe very basic, but it's just like try and be empathetic to the other person. And basically, how does Scott, how does Nick, , how do we wanna buy software?

[00:32:38] Scott: Mm-hmm. 

[00:32:38] Nicolas: And maybe you and I differ a little bit. We have obviously different stories And everything else, but like at the end of the day, , it's a discovery mechanism of like, are we good fit for each other? To help support achieve whatever objective. And so, the way we do it is actually that as like it's very human as a starting point.

[00:32:55] Scott: Mm-hmm. 

[00:32:55] Nicolas: Which I think we've been through sales cycle. I've been with my old companies through sales cycles of the NetSuites and the likes of the world. It's like literally like you can feel, you can smell the machine in the first 90 seconds of the call.

That's basically, you stare down some sort of like intricate like machine that you are about to get funneled through. And nobody wants to be treated like that. So like, flip it on its head. It's a very human-first approach. That means no three individual meetings to just get to the first person that can actually help answer any questions or show you the product.

Yeah. That is, put them on sandbox accounts as part of the sales experience where you can run your contracts and examples through to get a sense of if you can actually actually use the application. 

[00:33:32] Scott: Yeah. 

[00:33:32] Nicolas: That is fast communication mechanisms and back and forth via email, chat mediums that people use, text, phone, whatever it is.

[00:33:40] And so, it's very much like we're one team exploring together. And so it's a cultural human element, I think that is very different then how do we learn against that? Again, I don't think it's rocket science. It's just like the people that are in these conversations from our side are deeply empathetic with who we 

[00:33:55] Scott: Yeah.

[00:33:56] Nicolas: I don't even wanna say 'sell to', but who we work with basically. 

[00:33:59] Scott: Yeah. 

[00:33:59] Nicolas: And the people who we quote-unquote sell to slash work with, our accountants, controllers, our solutions consultants have literally been accountants and controllers of previous companies. They've bought NetSuite, they IPO'd businesses on softwares. They know what's important, and so it's just a conversation among peers.

[00:34:15] Scott:It's just like how does the customer get value from every single interaction that they have with you, even if they don't buy your software?

Even if they choose. So what, in our case, it's usually built. But it's like, how do you just like give them so much value that they're choosing a partner, they're choosing a long-term partner. And it works because your competition, in your case, it's like completely asymmetric.

It's like actually the value prop is like Nick showing up at their door or whatever, or someone else on your sale. Your market and our market, it's not the, like late-majority, optimize-every-little-last-edge. It's much more about that human consultant development. It's about how do we navigate this like, interestingly challenging moment for all of us and how do we get a partner who's gonna help us get through that.

[00:34:58] Nicolas: I was gonna latch on to one thing quickly because so important the partner to get through the next years.

That's what people are essentially looking for and for us, the contrast is so different to your legacy ERP system hasn't done any innovation in like, 20 years. 

[00:35:10] Scott: Yeah. 

[00:35:11] Nicolas: And then you have this newer company. We may only have 200 to 300 customers today, but still, like you have clear proof points that this thing works and you know, you have a partner that will be at the bleeding edge navigating whatever is to come... 

[00:35:24] Scott: exactly. 

[00:35:24] Nicolas: ...in the next couple years. I think that is so valuable for these teams. 

[00:35:27] Scott: Yeah. I was at a dinner about a year ago and there was like a bunch of public company CFOs there, and they were talking about how, I won't name the vendors, but they're talking about their ERP vendor license agreements and they're like, you know, because normally you're like, how do I get the minimum, like , the best price?

Am I gonna sign a five-year contract. And they're like, no. The only rule is you're only allowed to sign a one-year contract. It's like, I don't care about the price. It does not matter because this stuff is changing so fast that it's like, it's suicide to link into a five-year contract because like, I know what's gonna happen.

I know if some new AI technology's gonna come in. It's like in the general ledger. Actually if that's the level of these are traditionally very conservative buyers who are taking this fairly radical stance and it's like, as an entrepreneur, you're just like, wow, okay.

This change is like, is real. And so yeah, you must feel great. It's like opening up a huge door for you. For us, the way we experience it is it's like this AI is causing this generational turnover of business model. And that's causing all preconceptions to get questioned in this moment.

And so it's like okay, it's actually a weirdly, it's a jump ball even in the most stable parts of organizations. So, it's a very exciting time. I actually want to use that as a chance to talk a little bit about your business model. So, you know, broadly, how do you think about your business model?

How is it the same, different than like the folks you're competing with? Are you doing anything interesting there to kind of make the competition even more asymmetric? 

[00:36:56] Nicolas: Yeah, that's a part of the, I think, the business of [00:37:00] Rillet that I'm still trying to figure out. The pricing model is interesting and I don't think we're fully there yet in terms of how ERP accounting softwares in generally get priced because it's generally a cost saving, not revenue generating, value proposition.

There is some sort of revenue cash generation in there, but it's not the primary basically vector why people buy. It's usually the secondary. And so, for us, very specifically, one thing that we do quote-unquote Innovate on is we make it simpler. So some of these current systems are very well known for having deliberately obscure pricing models. And, sort of, to extract every single dollar that they can. So, I think that's maybe one counter move. Again, make it more human, more approachable, more understandable. That's one.

And then from then on, since we're on sort of a Metronome podcast here, talking about sort of outcomes-based and sort of usage-based pricing, that part is a little harder to grok, I think, for an ERP today.

The angle that I see where this is gonna change with agents becoming much, more and more prevalent. I do think you'll be in a position in say, two to three years, maybe not one to two years, but two to three years to really buy jobs to be done by like AR clerk or by staff accountant or by like certain functional titles in the organization be able to price against that. But I do think, to our discussion before, the workflows are not quite ready yet to, in earnest charge thousands and thousands of dollars for a certain workflow there. But I think that's sort of where it's gonna go. Less the usage, transactional based pricing that is very popular in other areas that you must see.

[00:38:35] Scott: Yeah, no, totally. It's like the agent stuff. Sometimes it goes to, well, very frequently it goes to, it depends on , how the agent value manifests. But, if it's like human augmentation, sometimes it's like, okay, cool, your team is now proportionally smaller because you're using our software, and therefore we're gonna charge you like a fee that kind of roughly scales with like, headcount savings that you're gonna get. That requires a very heavy proof of value kind of thing. And so, I suspect it would be workable with larger deals in your context where you're like, actually by the way, by using Rillet, you're able to actually save on headcount quite significantly.

The other area that, in kind of analogous systems to yours is if the insights and data out is actually like, it's like actually powering other types of workflows that can like sometimes be used in a more of a consumption or usage-based way. So it's almost like, the way I'm seeing it is like data science as a service.

So, it's like reporting is like, depending on how it's used. It can be monetized in a different way. And so that's like where sometimes you'll see like, a hybrid or consumption-based business model alongside of a platform fee or a headcount based fee.

[00:39:38] Nicolas: Can I ask you a quickly, actually, on that point, it's a very interesting thought. So, the monetization of the data or the data science, is sort of on the way out. How do you, or maybe your customers that you're advising around sort of pricing, how do you reconcile that within our specific cases?

Like, that's part of your core job, like the democratization of financial information, really being successful, living up to its fullest potential in the next five to 10 years for me means very democratized permission, and democratized access to financial data across an organization.

So you kinda don't wanna disincentivize that, but you obviously create a lot of value with it ? 

[00:40:13] Scott: It's an interesting question. I think, you know, we have a very similar belief, which is that basically like, you know, we are usage data married with spend data.

Okay. Like sales teams need that. Finance teams need that. Product teams need that. And engineering, you know, and customers need that. Everyone needs it, right? Okay. And we believe in that. But, we also bake it into our core pricing model that is like, look, the whole point is that we provide this to you in real-time at like very significant data scale.

And so we'll sell like essentially data egress and different forms to you, but we kind of make it part of the core package. Now, if you just wanna use our app and like access it that way, great. Like, it's not charged. And actually if you want to do like mass data export, even at low volumes, that's free. But then at very high volumes in real time, that's where it starts to become, and the way that we think about it actually is, it's like when you're small, we don't want to monetize on it.

We want you to like get it to everyone all the time, but as you scale and our costs go up, but also the value you get, like your team has grown exponentially. So, basically the way I would think about it, or the way we think about, it's like design a packaging such that like when you're smaller, you can kind of use it everywhere. And then when you're larger, you can still use it everywhere, but it's a core part of the value, and therefore it's monetized. And you have to basically attach it to some extreme value that they're getting.

For us it would be something like you're able to detect a potential churn of a customer , or a customer that should be upsold. Like, they're on a package where they're overspending or something like that. And so, it's like attach the value of the data egress to specific workloads, and then you monetize it that way.

But then design packaging such that, you can get ubiquity at the low end, or ubiquity at low access rate. You know, if it's like once a month, okay, who cares? Does that make sense? And then it also fits in this narrative of we wanna be a real- time system. And like our view of it is, the more teams that we're powering, the more value we're providing. Like our product value, you know, our costs should go up. Now, we don't want it to be like onerous or punitive, but that's how we've thought about it. 

[00:42:17] Nicolas: Yeah. That's super smart.

[00:42:18] Scott: Okay, cool. You have like 10 minutes. I wanna like dive into just kind of a little bit more about how you're building. Like, basically, how you're incorporating AI into your product and kind of the areas that you think are gonna like, drive the most radical change. 

So, talk a little bit about how you think AI is going to affect the finance, accounting functions of larger companies in the next couple years. What are the areas where , if someone's listening, you would encourage them to focus their attention so that they can see what's coming around the corner or maybe incrementally, months or quarters ahead of peer set?

[00:42:58] Nicolas: Yeah. Yeah. So the two bigger and mature customer set is number one, it is how do you parse through this, like, swaths of unstructured data at large scale? These are invoices, contracts, your shared service center somewhere out in, you name your country, Poland, India, whatever it may be. Like, what type of work do they do?

And is there an opportunity there to just using workflows, agentic is a fancy word for like LLM-related technologies to apply basically to these teams to just help automate and sort of, improve the throughput. So, that's usually also quite measurable.

And, it doesn't mean that you will lose these people. You can redeploy them into other areas. But I think that's one area that I find interesting at sort of at the highest level. And then at the most micro level, I'm always surprised by how many people talk about AI and want to use AI but themselves don't use like ChatGPT. Or themselves haven't actually tried solve some of the workflows themselves with these tools.

So very tactically, I had a customer that had a lockbox somewhere with like an insane amount of checks coming through. And, for any accountant out there who's done this process, they know what I'm talking about is basically, it's very hard sometimes to distinguish, depending on which lockbox provider you use, like what the total cash amount you're getting paid actually relates to in terms of underlying checks and invoices that are being paid.

And so you can run some of these files just through a a ChatGPT situation. And like basically code your own little like, whatever, what-is-it-called-GPT, to help through that process. And that was for me, such a awesome use case of someone who just had a problem.

You could have probably spent f in three months on like an RFP and like procurement and everything else versus just trying it yourself. So, that's sort of, I think, a great starting point because at the end of the day, our software vendors don't like to hear it, but like everybody's wrapping, OpenAI's technologies one way or another these days.

Again, the value in ERP is much larger than the LLM steps that we have as part of our product, but you can kind of mimic a lot of the functionality or at least parts of it with OpenAI directly. And so that's maybe the second piece- very tactically that is very also fun and exciting to play around with.

[00:45:11] Scott: Okay. Awesome. I also have to say like, really you're one of the most celebrated companies right now, at least in AI and enterprise software. So, congratulations. As one of the stars in the valley, how do you stay current with what's happening inside of Silicon Valley?

Like, things are moving so fast, like, what are some like information sources that you find that are particularly high bandwidth, either on like Twitter or X, whatever you wanna call it, LinkedIn? Like what are the areas, like, how are you staying plugged into the ecosystem of what's happening and changing as technology continues to advance?

[00:45:41] Nicolas: Yeah. This is so hard because also these sources keep on changing and then some things get hyped, and then, I don't know, fizzle out again. So, the two things that I do, and I wish I had, like, I don't have like one site or one podcast to point to, but the two things that I genuinely do myself, given all of us are strapped for bandwidth, is like number one, I do listen to certain podcasts or follow certain individuals and, and not necessarily directly read or interpret what they say, but like, sorry, I do that, but then I do interpret or like the subtext and how it applies to my domain. So there's certain business folks out there. And this is not only just the cool quote-unquote tech people. It's like, Jamie Dimon is, for example, an inspiring figure and individual the way he talks and thinks about the world.

There's some of these, like Wall Street professionals that have think literally 40-50 years of work experience and really interesting views on Take something. So, going direct via podcast to listen to these leaders directly on how they talk. And again, it's not about what they say, also how they say it.

How does that translate into maybe my leadership style and everything else? So, that's one thing on the podcast side. And then the other part is just source and inspiration from the team. 

So just listen to these folks, and maybe grab a coffee more versus less, even if it's just a 20-minute conversation. 

[00:46:49] Scott: Awesome. What's your favorite podcast right now? 

[00:46:52] Nicolas: I care more about the individual that's on the podcast.

Scott: Oh, okay. Cool. 

Nicolas: versus an actual series. I listen to some of the usual Valley podcasts, but like, certain individuals that I look up to that. Interesting. All right. 

[00:47:02] Scott: Well then, who's the person you're, like, you said Jamie Dimon, who's someone else that you're like tracking? 

[00:47:05] Nicolas: Someone else that I'm tracking is, he's a little bit of a maybe controversial figure, but Alex Karp? The way he communicates is very different from your average executive. And some things I wouldn't, I think take over into my style, but he is... he's different. Let's put it this way. And that's what makes him interesting.

[00:47:21] Scott: Totally. Yeah. He is definitely different. But, yeah, I agree. I actually find that folks like that, who clearly think differently, it just opens up my possibility space. Like, oh, okay. I would never say that, but like he's making a subtle point that actually is useful. For me to like incorporate , into the hive mind.

[00:47:34] , Awesome. And I guess one, one last question, which I think is, , is AI centric, but like what is. What is like your biggest personal workflow change in the past, let's say three months around ai? Like, you know, there's a lot of people here that are like always looking to learn from folks who are kind of at the bleeding edge.

[00:47:46] Like what's one thing that you've done that's like changed your workflow using AI recently? 

[00:47:52] Nicolas: Yeah, so for us, the thing that impacts me personally the most is video call recordings. So just gotten so [00:48:00] efficient. We use certain tools for sales calls, customer calls, to summarize query information, and that way as a CEO you can literally access hundreds of parallel conversations that are going on in a very targeted fashion. Sample coach, I would listen on a run or like, when I'm going through, maybe not quite going through security, but I'm like walking in an airport somewhere, I would listen to a sales call that a new AE had or a person.

[00:48:24] So that's sort of like completely transformed my access to what's going in the company while being able to give to the organization a lot of coaching and training, , and questioning, in a good way, questioning, like challenging, of individuals which has, I think, helped me personally a lot and hopefully helps the organization a lot.

[00:48:41] Scott: That's awesome. I love that. Awesome. Well, I just wanna say thank you. I think, this has been great. Like, you have a ton of interesting things to say and obviously Rillet's an incredible business, and I hope we get to work together more in the future. It was great, talking with you Nick, and I hope you have a great afternoon.

[00:48:56] Nicolas: Yeah, likewise Scott. Thanks for having me. 

[00:48:58] Scott: Sweet. Thanks. [00:49:00] Bye.

[00:49:04] Thanks for tuning into this episode of Unpack Pricing. If you enjoyed it, we really appreciate you sharing it with a friend. We'd also love to hear from you. Feel free to email me@scottatmetronome.com with feedback and suggestions for who you'd like to see on the Future Podcast.