John Glasgow of Campfire, chief executive officer and co-founder, appears in this theCUBE Research interview hosted by John Furrier and Gabe Olave at the NYSE Wired Business Transformation Edge series. Glasgow discusses artificial intelligence-native finance transformation, enterprise resource planning migration, domain-specific models for accounting workloads and agent-driven finance automation. They explain Campfire's approach to migrating legacy ERP systems, building post-trained finance models and deploying agents to automate reconciliations, accruals and continuous close workflows.
Key takeaways include accelerated migration timelines, deterministic finance models and agent-driven automation that reduce manual accounting work. Glasgow states Campfire's post-trained models and agent layer enable near-zero tolerance for inaccuracies in financial reports, support real-time continuous close and convert token-driven AI costs into manageable capital expenditure when running proprietary models. Analysts on theCUBE note parallels with Ramp and a broader shift up the AI-native stack.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: C-Suite Business Transformation. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: C-Suite Business Transformation.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for theCUBE + NYSE Wired: C-Suite Business Transformation
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You are already logged into TheCUBE Network as
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for theCUBE + NYSE Wired: C-Suite Business Transformation.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: C-Suite Business Transformation. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: C-Suite Business Transformation.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: C-Suite Business Transformation
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: C-Suite Business Transformation. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
John Glasgow, Campfire
John Glasgow of Campfire, chief executive officer and co-founder, appears in this theCUBE Research interview hosted by John Furrier and Gabe Olave at the NYSE Wired Business Transformation Edge series. Glasgow discusses artificial intelligence-native finance transformation, enterprise resource planning migration, domain-specific models for accounting workloads and agent-driven finance automation. They explain Campfire's approach to migrating legacy ERP systems, building post-trained finance models and deploying agents to automate reconciliations, accruals and continuous close workflows.
Key takeaways include accelerated migration timelines, deterministic finance models and agent-driven automation that reduce manual accounting work. Glasgow states Campfire's post-trained models and agent layer enable near-zero tolerance for inaccuracies in financial reports, support real-time continuous close and convert token-driven AI costs into manageable capital expenditure when running proprietary models. Analysts on theCUBE note parallels with Ramp and a broader shift up the AI-native stack.
play_circle_outlineAI-Native Finance: Campfire Model Automating Accruals and Reconciliations with Deterministic Python/SQL for Big Four Audit Compliance
replyShare Clip
play_circle_outlineScalability: AI ingesting millions of rows versus legacy ERP limits.
replyShare Clip
play_circle_outlineIntegration ecosystem: payroll and other systems via APIs.
replyShare Clip
play_circle_outlineZero-code, self-serve deployment and data migration tooling.
In this interview from theCUBE + NYSE Wired: Business Transformation Edge, John Glasgow, chief executive officer and co-founder of Campfire, joins theCUBE's John Furrier to discuss the great migration of finance teams from legacy ERP systems into the AI-native era. Glasgow explains how Campfire built its own post-trained foundation model purely for finance and accounting workloads, shifting outputs from probabilistic to deterministic by having the model write Python and SQL directly, a move that meets the zero-tolerance-for-error bar required for Big Four aud...Read more
exploreKeep Exploring
What was the talk about, and what is the speaker's take on ERP systems?add
How does your solution leverage AI and agents to process large volumes of ERP data and automate the long tail of finance and accounting tasks, and what business benefits does that deliver?add
Does the company provide payroll services?add
What is the ideal way for a growing company (e.g., $50M–$100M) to deploy and migrate to Campfire, and how much engineering/coding effort is required?add
>> . I'm your host of theCUBE here at theCUBE's NYSE studio. Of course, we have our Palo Alto studio connecting Silicon Valley to Wall Street. This is our Business Transformation Edge series where we talk to the leaders who are making it happen in the AI native era as intelligence gets injected into the business. The rules of engagement, the rules of systems change. We're seeing that with NVIDIA up and down the stack. This is the number one conference. It's an opportunity for business model transitions to the new normal. That is AI native. John Glasgow is here, CEO and co-founder of Campfire, a fast -growing company doing the great migration with ERP. Great to see you again. Thanks for coming in.
John Glasgow
>> John, thanks for having me on today. Really excited to be here.
John Furrier
>> So you're in the building for a Jefferies investor conference. We appreciate you taking the time. We saw each other here in New York about two months ago on a panel, and we were riffing on this idea that, okay, ERP systems can be built up.
John Glasgow
>> Yes.
John Furrier
>> They're hard. They're old. They're antiquated. Some would say, well, they've got the system of record in there, so why don't we just build on top of it? You have a different approach. You see something different. You gave a talk today downstairs here at the NYSE.
John Glasgow
>> Yeah.
John Furrier
>> What was the talk? What's your take on ERP?
John Glasgow
>> It was us and Ramp, and it was called The Great Migration. and it's all about there's this huge wave of folks that if you think about there was on-premise to cloud and now there's cloud to AI and folks are ready to move into the AI era. I'd say engineering was probably the first to go and then we were all using it for personal uses and checking on things at home about cooking recipes and now models have really gotten to a point where they're very good at finance and accounting workloads. And so, us and Ramp are talking about how there's just a groundswell of folks coming over that are really ready to go AI-enabled.
John Furrier
>> What's driving the migration? Models are getting better, check the box there, but models are separating from the systems that actually they run on, so seeing that kind of separation, they still work together. What's going on at the systems level? Because the success of Ramp and your company, Campfire, is showing a template of how you could actually go in to an existing function and just do it a little bit differently with the modern platform and tooling, and certainly autonomous is here with agents. What's different?
John Glasgow
>> Yeah, I would say the first one is the way to migrate, and that was part of the talk as well. We've gotten fundamentally better at moving folks off of legacy ERPs. And so the time to make a move has gone way down, the value has gone way up. And then the other big one is we have our own post-trained foundation model that is purely trained for finance and accounting workloads. And so a year ago, people were saying, hey, this isn't very accurate. And obviously, in finance and accounting, if the data is not accurate, it's really not usable at all. Tolerance for error is close to zero, zero for many use cases. And so we're seeing, we have trained our model to go from probabilistic, where we're feeding all the financial data into the model, to now deterministic, where it's actually writing Python and SQL. and so once you confirm the accuracy, we can build custom financial reports, we can do all sorts of accruals or reconciliations with AI, but still introduce it into a Big Four audit environment, still get incredible accuracy, and so folks are saying for the first time, AI's really ready for me.
John Furrier
>> it's interesting, you pointed to a trend we're seeing from a business model standpoint where, how do I say this, maybe I'll use a cloud metaphor, taking primitives and moving up the stack, and the adaptiveness of the platform becomes the key because you said the data's there, you've got to be right. And robotics, by the way, is having a very similar experience. Safety, you can't be wrong. Zero tolerance for error.
John Glasgow
>> Gets dangerous.
John Furrier
>> In finance, same thing. So when you have that ability to create that autonomous layer, self -service becomes easier. Execution. So, yeah, they know their models, they know what they want to do in their apps, but now you have a platform. What's the dynamic there? Why is it being successful now? Why is this happening? I see AI native companies, say I'm a startup, I'm growing at escape velocity, I'll use Campfire, I'll buy an ERP that's AI native with you guys, but as companies start to migrate, what makes that possible? Is it the tech stack? Is it the, explain.
John Glasgow
>> Yeah, of course. Well, we start with the scalable core. And so legacy ERPs can't handle a lot of data. Now that was fine when it was just humans looking at the data, AI can consume millions, hundreds of millions of rows of data in ways that just we cannot. And so the first one was like, we pull in every row of data for our customers, and where previously that was not the case, and then we put our own model and our own agent layer, and just building agents on your own. For the long tail of tasks, I, in the early days of Campfire, met with so many finance and accounting folks manually creating invoices, manually doing reconciliations, and it was because rule -based logic was not able to really help them for this long tail of work, agents can. And so we offload the kind of day -to -day tasks to agents, and now they're much more strategic in the work they're doing every day. Now they're helping grow revenue, improve profitability, drive efficiency.
John Furrier
>> I hear the edge case scenario all the time. And back in the old days, it would be like, oh, it's just an edge case. And so rules -based systems used to be okay, but there were plenty of edge cases, more edge cases than people were documenting. Talk about that dynamic and where in the long tail the agents solve that problem because with AI you don't need to code rules. You can have a graph, you can have a context graph, you can get all kinds of domain specific specialized intelligence that you're doing.
John Glasgow
>> Yeah, I think one very tactical example is accruals, it's a cost estimate. And there's a lot, particularly in an audit environment, the more accruals you're doing it becomes just a huge amount of manual work. It typically is done by doing some math in an Excel file, and then you're gonna manually create a journal entry. And it's something that you can't offload to rules most of the time, and maybe you're even chasing a law firm to figure out what the bill is going to be. And so AI can do a lot of this piecing together, a lot of this, call it less science, more art, in terms of figuring out what the accrual amount should be. And then we're seeing agents just kind of take over this. each company is going to do accruals a little differently, each vendor is a little different, that's very unique to the business. And that's where AI can just take on a lot of these unique business needs.
John Furrier
>> Because they can actually take the domain expertise, the undocumented knowledge, and apply it.
John Furrier
>> Correct.
John Furrier
>> What about other CFOs that are out there right now? One of the people watching this program, a lot of CFOs and HR departments are thinking, okay, you have agents become workers, Productivity, that's revenue. Efficiency, that's cost savings on the ERP license, and then maybe new capabilities, but they're crossing a lot of boundaries operationally. So it's easy to take a CFO and make them a COO, harder to take a COO and make them a CFO. So you're seeing the role of the CFO become very operational.
John Glasgow
>> Correct.
John Furrier
>> What's your take on that, and what advice would you give to folks watching who are really struggling, and not struggling, wrestling with decisions on how to approach AI, how do I instrument it, How do I deploy it foundationally?
John Glasgow
>> Yeah, I think particularly for folks, AI is fundamentally shifting when you want to see the data. So an example, typically people would look at, they would do the monthly close, and then they would look at the data. Many of our customers now have large compute costs, large AI costs, that you can quickly go negative contribution margin through a month if you're not very thoughtful because models can get out of control. And that's particularly on the OpEx side. We saw Uber famously announce.
John Furrier
>> You mean not compute, token costs.
John Glasgow
>> Token costs. Yeah. We saw famously that Uber in Q1 announced they went through their entire budget for the year on token costs.
John Furrier
>> By the way, that's consistent with many companies.
John Glasgow
>> Yeah. That's not a one-off. They're not alone. They were just the only one kind enough to acknowledge it publicly. And so folks are saying we need real -time visibility. And they come to us and say looking at the end of the month is no longer going to work. And so we're able to really help them whether it's on the customer side, contribution margin, or internally, we can give the visibility. And previously, things like business unit and standing up a lot of custom dimensions was easy to create but very hard to maintain. With AI, they can do a massive amount of maintenance in terms of labeling data, and so that's been a big unlock on Insights.
John Furrier
>> I love how companies are adopting, especially in the area that you're in, because there's so many great wins. you can handle accruals, you can come in other parts of the business, just expense management, you've got finance, cash flow, all these different pieces. You mentioned closing the books. you're now looking at real-time data. You can actually have visibility into the close before the close.
John Furrier
>> Yeah.
John Furrier
>> And with the intelligence, run scenarios on many different things. Is that really new, or is AI enabling that? Because I haven't heard a lot of that in the past.
John Glasgow
>> Yeah. Yeah, I would say this concept of the continuous close is newer. And the reason was before there was just too much work to do through the month. And so once a month was kind of the limit for an accounting team. Now with AI, you can move into more of a real-time close.
John Furrier
>> brute force is one way to do it.
John Glasgow
>> Sure.
John Furrier
>> Throw bodies at it, right? That doesn't scale. Talk about your customers. Take me through some use cases because I find a lot of people ask me all the time about this area. Hey, I'm on this system. I won't name names, but we all know what they are. Or, hey, I'm a fast -growing company. We just started last year. We've outgrown QuickBooks. I want to do P &Ls. I have just bespoke, just fragmented systems. What do I do? Or, hey, I'm a big company. I want to modernize. There's many different kinds of orientations to what Campfire offers. What's the use case? What are some pro tips? Give some examples of how people engage and should think about it.
John Glasgow
>> Look, there is a huge mandate right now from executives and boards saying, you need to go AI native for each function. In engineering and sales, there's some great tooling out there that they've all made big leaps, like Cursor and Clay. And the finance teams are saying, there's really been nothing for me. And so they see us and they say, wow, this is truly the Clay. And they're like, we don't have any coders.
John Furrier
>> Where's the staff to code? We're not going to hire coders? That's a question too.
John Glasgow
>> Exactly. And so one of our customers, Replit, has grown 20x revenue without adding folks to the accounting team. And so they're seeing tremendous value. But also businesses, whether it's professional services, everybody from investment, just at the Jefferies Conference, everyone from investment banks over to nonprofits are coming to us and saying, we are looking to make an AI transformation. And so it's been amazing to partner.
John Furrier
>> You know, John, when I was a kid, there was an amusement park nearby, and I always wanted to ride the roller coaster. And I always measured myself on the line. You've got to be this tall to ride. Finally, you get there, you ride. Is there a line for customers to ride Campfire? Is there a threshold of table stakes, requirements, prerequisites? Sure. Do you have to have? Can I come in with a sprawling QuickBooks mess with bespoke spreadsheets and a TriNet account or whatever outsourced system?
John Glasgow
>> We've seen it all, from oil and gas ERPs to 100 different QuickBooks accounts.
John Furrier
>> But can a small guy get in there?
John Glasgow
>> A small guy can get in there, yes. You can get in. And we've had people start their company on Campfire all the way through. We have public companies on Campfire into the enterprise as well. We have customers headquartered in every continent today. We have offices globally.
John Furrier
>> Do you guys do payroll?
John Glasgow
>> We don't do payroll.
John Furrier
>> What's your experience with payroll? Because that's always kind of like the DMV of business. No one likes it. They have to use it. Maybe that's a bad way. I shouldn't say that. But there's a lot of money going through there. Service is terrible among almost all the big ones that we know.
John Glasgow
>> Sure.Yeah. We're not in payroll. We integrate with all of them. We have 200 payroll integrations. So if you do find one you like, we would love to integrate it and make sure it's seamless is our focus.
John Furrier
>> And what's the integration? Just APIs or just?
John Glasgow
>> Yeah. We've got direct integrations on the payroll side.
John Furrier
>> What's the ideal deployment? If you had someone say, hey, John, I want to go. Let's just say I was a $100 million company or $50 million company, I'm growing. I want to just restart and go with Campfire. What do I do?
John Glasgow
>> Yeah, I would say obviously give us a call. But I think really anybody, irrespective of where you are. So some folks are on an enterprise ERP, and they come to us and say, hey, we really want to transform. And for some folks it's day one. But if you have any accounting data or none at all, we have a team that will kind of help transform it. And we have some tooling that will really automate the whole process for you. And even before signing up, we can actually migrate all of your data for you. And you can do a live demo with all your own data and it's quite magical.
John Furrier
>> How much coding would I need to use Campfire from an engineering standpoint?
John Glasgow
>> What's the... Zero coding. So all self-service. All self-serve. If you want to set up custom dimensions, you want to add a subsidiary, it's all self-serve for the end user.
John Furrier
>> Okay, so here's a scenario for you. I see this a lot in some of my interviews with founders is that product market fit is my focus. My inside out view is get my product to match the outside in with the customer, but the market's shifting. I had a P &L today that's not going to look like this, so there's a lot of agility needed on the business logic side, essentially finance, but then mapping process. Do you see that as an issue, and does that come up a lot, and how would you talk to the customer who says, hey, I want to make sure I'm outside in, matching the value proposition to what my customer will pay, and then making sure all my backend processes lock in behind it.
John Glasgow
>> We see that all the time. So one great example is everybody's moving from subscription to, I remember when everyone moved to subscription billing, and everyone's moving from subscription billing. And so we're told we're the deepest on usage -based consumption. And so the reason why you have to move away from subscription, of course, is AI. The costs to deliver are so high. It's almost like the margin is like hardware, right? And so you get into a lower margin profile, and so you need to align value and price on token consumption. And so we've got a usage -based billing engine that we've really adopted and built for folks that are now adding that.
John Furrier
>> Do you see a financial metric where the token maxing turns into value maxing, meaning back in the day when I was in college, you were judged by your code by how tight it was, smaller the better.
John Glasgow
>> Sure, yeah. Because memory was tight back then. Still is.
John Furrier
>> Still is. Tokens, people were blowing their tokens. It's the leaderboard. All the tokens I used, that was more like just not really efficient. So I think there's a swing back to token efficiency to value. I got this outcome from this token size. Smaller the better. You get maybe more compute. To kind of use Jensen's analogy, I would say that's a better model for finance. We say, hey, I only use X number of tokens, but it's just massive output. That's better value.
John Glasgow
>> It absolutely is. We never had token maxing internally. I never fully understood the concept. We're all about like employees, you can use unlimited tokens. There's no token caps. But it's really like make sure you're getting value out of the token. One thing that's been big for us is the rise of open source and kind of like us taking models and retraining them. And then from a cost perspective, you no longer have tokens. It's purely CapEx. We have our own GPUs, at Campfire. So we run our own model.
John Furrier
>> You have a token factory internally.
John Glasgow
>> Yeah, we do. And so from a consumption perspective, it's no longer cost of goods sold, it's now a capex piece.
John Furrier
>> How did you get your models so trained and accurate for your domain? Just give us a taste of the scope of the complexity or the job it took to do that.
John Glasgow
>> Yeah, well fortunately for us, accountants are meticulous about labeling data. That's a good thing. Accuracy is critical, and so we've got a lot of very clean data. And AI engineers come to Campfire from other categories like solving coding items on the healthcare side. And it's very messy data.
John Furrier
>> Yeah, so they have dirty data and you have clean data.
John Glasgow
>> A lot of dirty data. They just love it. They do love clean data. And so we're in a unique spot. And the funny thing is nobody's posting their general ledger online. Like you look at law, you look at coding, there's tons of data online for the models.
John Furrier
>> You mean the public domain.
John Glasgow
>> Public domain. Yeah. Publicly available. There's a lot of code and law on the internet. Nobody's posting their general ledger on the internet. So we of course have opted in with customer consent, but the value is driven back to them in that they see the value when they opt in for training our model.
John Furrier
>> All right, well, really appreciate you coming in. This is probably one of many more conversations. It's a hot area. You guys are doing great work, AI Native. I guess my question is here at the investor conference, how'd it go? Can you share some momentum on the company? Milestone, size, funding, cash in the bank. How many customers do you have? What's your revenue? Yeah.
John Glasgow
>> It's a great question. I just got all those questions about an hour ago. Well, look, we're three years old. We raised $100 million in funding, primarily from Y Combinator and Accel and Ribbit Capital. One of the big questions was, there's so much enthusiasm from the venture community. This is the largest category in application software. They're when are you raising more? And so that was the top question I got today. but also many of them see in their portfolios that Campfire is the number one fastest -growing product in the CFO suite for their portfolios. And they say, wow, you guys are doing a phenomenal job in driving efficiency and improving reporting.
John Furrier
>> You're on their radar by default.
John Glasgow
>> And so we're on the radar that they're hearing great things, and that was obviously great for a founder to hear. We're 150 employees. We're in three offices globally.
John Furrier
>> Are you guys sharing revenue numbers at all?
John Glasgow
>> We're not sharing, but we did 10x revenue year over year. and we are on a similar trajectory over the next year as well.
John Furrier
>> How do you compare to Ramp? Because they do share some of their numbers out there. How would you compare yourself to, say, Ramp?
John Glasgow
>> Ramp's a bit older and they're a bit larger. So we go to market with Ramp. We're very close. We don't compete.
John Furrier
>> You partner with them.
John Glasgow
>> Yeah. They are a great partner of ours in that we're the general ledger and they do the spend for customers. Yeah, Ramp's an awesome company.
John Furrier
>> Well, congratulations. Thanks for coming in. I really appreciate it. Business model transformation starts with, one, getting the business model right, managing it, measuring it, getting that general ledger, but also the new platform of how business operates is not just financial, it's going to be in our pockets, on our phones, voice activation. Agents will make things easier, but accuracy still will be the number one thing, and that's certainly seen a lot in the business model. We're still doing our part here on theCUBE. Thanks for watching.
John Furrier