AI Agents Are Changing Accounting Forever | The New Future of Finance & ERP with John Glasgow

What if the biggest bottleneck in finance isn't the software, but the way work gets done?

In this episode Chris sits down with John Glasgow, Founder and CEO of Campfire, an AI-native ERP platform built for finance and accounting teams. John explains why simply adding AI to legacy systems falls short, how agentic workflows are changing accounting operations, and why finance leaders should think of AI as a teammate rather than a tool. He shares real examples of AI handling reconciliations, treasury management, reporting, and close processes while keeping humans accountable for outcomes.

The conversation explores practical adoption strategies, how finance teams can build trust in AI, and where human judgment remains essential. Leaders will come away with a clear framework for introducing AI into critical business processes while maintaining accuracy, accountability, and stakeholder confidence.

Chapters:

(00:00) Introduction
(02:00) Meet John Glasgow and Campfire
(03:00) Why AI-First ERP Is Different
(06:04) From System of Record to System of Work
(08:26) AI as Operational Capacity, Not Just Productivity
(13:47) Inside Campfire's Ember Agents
(16:04) Building Trust in AI for Finance Teams
(18:49) Treat AI Like a New Hire
(22:13) Where Humans Must Stay in the Loop
(28:49) Customer Results and Real World Impact
(31:19) Does AI Replace Finance Jobs?
(34:27) The Skills Future Finance Leaders Need

Resources:

🔎 Find Out More About John Glasgow:

John Glasgow LinkedIn
https://www.linkedin.com/in/johnglasgow/ 

Campfire 
https://www.campfire.ai

Campfire LinkedIn
https://www.linkedin.com/company/meetcampfire 

🛠 AI Tools and Resources Mentioned:

Campfire 
https://www.campfire.ai

Claude 
https://claude.ai

Anthropic 
https://www.anthropic.com

NetSuite 
https://www.netsuite.com

SAP 
https://www.sap.com

Workday 
https://www.workday.com

QuickBooks
https://quickbooks.intuit.com

Xero
https://www.xero.com

Transcript

Copy Transcript Copied

SPEAKER_01 0:35

God forbid you get something critical wrong, then AI can be incredibly destructive.

SPEAKER_02 0:39

There is a lot of excitement and enthusiasm about what's possible, but not a lot of trust in the system.

SPEAKER_01 0:46

I had a 15-year finance career, so me trusting AI and hallucinations and I get all of that. I think within finance, the easiest place to start is have it review your work. You don't need to start work like it's running autonomously on my data.

SPEAKER_02 1:00

Would you fire that person on day one if you gave them crap instructions and they gave you crap output?

SPEAKER_01 1:05

Absolutely not. Just treat it like a new hire on the team. Learn how to work with it, learn where it's good, don't trust it in the beginning.

SPEAKER_02 1:12

What are some areas within finance and maybe at large that you think humans still need to stay firmly in the loop?

SPEAKER_01 1:18

Humans need to be accountable for the output of their AI. Own the output and be accountable for the output and be able to speak to the output as if it's your own. Otherwise, like AI slop is introduced into the system, and then you personally kind of lose trust of your stakeholders.

SPEAKER_02 1:33

Welcome to Using AI at work. I'm your host, Chris Dag. Each week we'll be learning how today's business owners, entrepreneurs, and ambitious professionals are getting more done with smart use of tomorrow's tech. Let's get started. Right now, every business leader is asking the same question. What are we going to do about AI? If this is you, ChiefAIOfficer.com has the answer. We give you a simple path forward where we provide executive and team training so your people know exactly how to safely use generative AI in their day-to-day. We also manage the deployment and implementation to make sure tools actually get adopted and deliver results. And we'll also guide company-wide transformation so AI becomes part of your operating system, not just another shiny object. The companies that act now will increase productivity, cut costs, and grow faster than their competitors. Those that wait will get left behind. So if you want to make AI work in your business, visit ChiefAIOfficer.com and see how we're helping companies of all sizes finally get results from AI. Hey everybody, welcome to another episode of Using AI at Work. My name is Chris Daigle, and I'm the host. And today our guest is John Glasgow. He's the founder and CEO of Campfire, based there in the San Francisco area. It's an AI-native ERP platform focused primarily on finance and accounting workflows. So the ambitions at Campfire are grand, and they plan on being a replacement for legacy systems like NetSuite, especially for companies that understand they need AI embedded directly into their core finance operations. So John, I've got a lot of questions for you today. But before we get started, is there anything as by way of intro that you think might um help our audience uh kind of understand, I don't know, your paradigm and the way that you're looking at things over there at Campfire?

SPEAKER_01 3:19

Um it's a great question. And Chris, thanks for having me on. I think the only thing I'd add there, you I mean, look, you obviously described it well, so I'm not gonna add much to that. I think the only thing I'd add is folks are saying, like, how do I rethink workflows from the ground up? And so like old systems with AI slapped on top, it hasn't just not worked. And so it's like, how do we like fundamentally rethink everything with AI first? And so like when we say AI native, it's like we have our own foundation model that's like post-trained. We've got our own agent layer and our own ERP, and we've collapsed it all into one. And then we said, what if we started over? Because we started the company after ChatGPT 3 came out. So we started everything over from scratch. What would you do differently on every workflow if you took all that and put it together? And that's Campfire. And so customers like Replit or PostHog or customers, and they've come to us and said, we just want to fundamentally rethink how finance and accounting operates, and we've been um honored to support them on that journey.

SPEAKER_02 4:17

I don't want to uh uh put any pressure on you, but when you talk about these attempts by maybe some of the incumbents to just add AI into either the narrative, whether or not it was actually part of the product, are there any of those that have tried it that you think they just didn't pull it off? And if you feel comfortable naming them, I mean we're not we're not crap talking them, but I think it's good context for a listener because they may not know. They may say, oh, no, no, no, what we're using, they use AI.

SPEAKER_01 4:44

Yeah, I think all the incumbents have essentially said our AI strategy, here's an MCP, and go go add Claude or ChatGPT on your own and go run it third party. The big problem there is a lot of folks say, like, I want the AI running within my ERP, whether it's for permissioning or controls or just like having it generate reports for me at scale, like having it ingest millions of rows of data. There's just a lot of like reasons to have it built into the system. In addition to like when you go and perform a task in Campfire, the AI is like fused into the task itself. So reconciling cash, like the the incumbents don't have this, it'll actually like use AI to automatically perform the task. And if the AI can't perform the task, it'll pull you in, and the human in the loop is what it's called. It'll pull you in when it gets stuck. And so like the historical systems have used rule-based automation or OCR, which is more of like logic-based, like if this then that, where of course you've seen the power of gen AI, incredibly powerful. So the long tail of subjectivity that maybe a rule doesn't work for, AI can be incredibly powerful. And when it's infused into the task itself, then uh you get very different results.

SPEAKER_02 6:07

So let's let's actually talk about that. Um this concept of taking, you know, from from ERP as a system of record to ERP as a system of work, where traditional ERPs are storing the financial data, but the humans are st the humans are still working with the data, right? Around the reconciliations that you talked about, the clothes prep, all that stuff. But in Campfire's um functionality, the ERP itself is doing more of that work using your AI agents. Is that accurate?

SPEAKER_01 6:41

Correct. So here's a here's a very real example for you. Is I used to be in finance at Adobe, and we would do a treasury roll forward, um, and there's a lot of things we would manually do, like daily, weekly, monthly, quarterly. A lot of things we would just like rerun. So like maybe there's a a weekly finance team stand-up, and you would update all the numbers in the slide by hand, and you would kind of share it out, and you would write some commentary by hand. And a campfire, like we built a treasury agent, and so you know, treasury pays more than like a checking account, as you know, like 4% for zero, for example. Um, we have a daily agent that takes via MCP, it looks into all of our systems, like our corporate card provider and our bank accounts, and of course into Campfire. And every single day it does like a five-day forward-looking view, and it's like, hey, move cash from you know treasury over to operating or back and forth to optimize yield. And it sends our team a Slack whenever we need to move cash. And you can automate that, but we've decided to just not have it go into our bank account and move money. Um it's picked up another 50k a year in yield by just helping us optimize cash. So I think then you take that across hundreds of finance tasks and hundreds of accounting tasks, like that weekly DAC that can't AI update the numbers in the DAC, maybe even more accurate than a human. Um so a lot of this work is just being automated. But I just um I just asked the agents the other day, like, generate five like things I should do differently as a CFO at campfire. And three of them I went and executed on. And when I was at Adobe, like we had hundreds and hundreds of people on the team, and like AI can comb through 10 million rows of data. Um, it can look across systems, it can look at you know things that like is just uh uncomprehensible for a human to do. And so now like the ability to just like comb through vast data sets and find insights that a human would be not able to do is is actually quite compelling. And so you don't need to have like an army of people like we had at Adobe to get in tremendously valuable insights.

SPEAKER_02 9:01

So not everybody listening to this is is in the finance function. Um but I I I want to ask this question through the lens of if I'm if I'm not a finance professional or I'm not in you know in the finance department, but I I'm what I'm hearing, because already I'm like this sounds like a much better way way to um uh leverage technology within our financial department. If AI can give me the suggestions through you know with a 60% accuracy as far as like, hey, these are good ideas, or uh uh generate another $50,000 in yield dynamically without the humans having to necessarily be doing all the thinking, just the approval side of thing, like that sounds very compelling. So, um, how should a CFO think about replacing what they've what they're already on versus augmenting what they're already on in a situation like this? Is it possible if I'm on one of the the the Net Suites or one of the legacy ERPs, what would a transition look like? And and what would be the criteria for me to know, hey, this is worth it?

SPEAKER_01 10:04

Well, certainly like we would love to chat and go deep on any specific situation. I think for um the folks that we meet with, I mean, a third of our customers are coming from Net Suite itself. The rest are various different systems. So, including we've done SAP and we've done Workday and we've gone into the enterprise too. Many are coming from QuickBooks or Zero on the small business accounting software side. But ultimately, whether it's small business or enterprise, they fundamentally say we need to adopt AI, our kind of finance function, just call it GNA, you know, general and admin, which is inclusive of finance, legal, HR ops. That's an area that is not getting a lot of headcount in the current environment, but they're being asked to do more with less or more with the same. You know, the the business is growing fat. You know, Replit has grown tremendously 20x revenue on Campfire without adding accounting headcount. And so I think, you know, everybody is being our our recruiting team came to me and said, you know, we've added 100 hires at Campfire in seven months, and we've added one recruiter in that time period. And they they came to me and said, we want to buy an agentic system to help us systematically find high-quality candidates, better than a human could do, and then process them and then help with the end-to-end funnel. And we've seen just tremendous leverage for our very lean um recruiting team. And so I think it's it's beyond just finance and accounting. I mean, as the founder here, I went to the whole company and I said, once Claude Cowork came out, I went to the whole company in an all hands and showed one thing I had built for every single department. So for our implementation team, I showed like, here's like something you can do for a sales team. I said, here's a custom deck for this that that Claude built for me, for the finance team, for literally everybody, and said, if you don't try something with AI first, then we have a problem. And and showed them all, all the power and gave everybody a license. And so whether it's like purpose-built tooling like Campfire or just generic, like help me write an email reply to a sales prospect, like fundamentally, I believe AI should be taking a first pass at everything. I mean, we had a board meeting this week and AI prepared, I dropped our board, you know, AI prepared the board deck for me, but then I fed the board deck into AI and said, help me write the narrative. And and then I, you know, obviously I process and synthesize what to say. Sure. But ultimately, like whether it's just like categorizing a bank transaction or like board level narrative, I fundamentally believe everybody, including myself, needs to be AI native.

SPEAKER_02 12:49

And I like that you uh did an all hands when Claude Cowork came out. So, listeners, if you're dabbling with anything and you haven't shared that uh those wins with the rest of your team, I would encourage you to do that. The more that the rest of your team sees each other using these tools, the faster some formal approach to, hey, how do we become an AI emergent organization is going to happen. So this is interesting because 3.5 comes out, 4.0 comes out, whatever. The conversation was about AI as a productivity assistant, right? The human's still doing the thing, but they're able to write the email faster. But what we're talking about here in that in that recruiting example and the the um the growth and the demands of replit's you know, uh financial operations without the headcount expanding, is showing us that that AI is becoming operational capacity, not just a productivity assistant, right? So and I know that you guys, you've got agents built in embers, right? Ember agents. Uh-huh. Okay. So what would be the difference between something that that I would say a lot of people, especially if they're on Microsoft, are probably using Copilot to some degree as that productivity assistant in the finance environment as compared to an Ember agent, like bringing that operational capacity? I guess help the the listener who well, are are people already using AI and finance, understand the difference between what they're probably doing now and what uh an AI agent, an ember agent from Campfire's ecosystem would look like.

SPEAKER_01 14:22

Yeah. Well, I'd say the first one is they are built directly into the workflow. And so it's like assisting you through the process. Cowork is often more of like a sidebar or like a browser, you know, something in the in the header. This is like it's assisting the human through the task itself. So it's like the human, it's it's like an employee on your team, and it's literally performing the work with you in the actual workflow. So a fundamental example, reconciling cash is like an accounting task. And if you're on cowork, you might like download cash from a bank feed and upload it to cowork, and then you'll like ask cowork to categorize it, and then you'll like upload it into your ERP. In Campfire, it's just like doing the work for you in the ERP. It has all the permissioning and auditability of so like if you're audited or if you want someone to review it, that's all built into it. And then it'll just like raise its hand when it when it gets stuck. Like if it's like, hey, I'm not 100% confident, and we can talk about hallucinations too, but ultimately like it'll then work with the human through the task as opposed to like uh patching systems together. I think the other one is actually quality of output. We have seen because we have our own foundation model, our own agents, and the own ERP built into one. The you know, obviously, accuracy is critical in finance and accounting for taxes or for any sort of reporting that you do. We have seen a higher quality of output. So you, yes, you can go to cowork. Um, but I think the the other one is my comment on it is coming to you and raising its hand on things you didn't even know about. So we have agents that are combing through all of the data in your account, and it'll tell you like, hey, this actually looks like it's wrong. Or hey, you might want to like take a look at this anthropic bill, looks like you double paid it. It's just like constantly, like the treasury example I gave, like cowork, there's no way of really uh really doing that at cowork unless you do a heavy build yourself. Um and so I there's there's a lot of value beyond just like hey, it's built in. It's it's really about the quality of output and the ease ease of use.

SPEAKER_02 17:30

So we we do like we go into companies, we educate them on it, we identify pilots, and we uh work with them on deploying those pilots. And finance, obviously, as you already mentioned, like accuracy is expected. And it can be a situation where uh there is a lot of excitement and enthusiasm about what's possible, but not a lot of trust in the system. Because they because they they haven't had the exposure, and it's it's mainly like I I just don't know enough about it to trust it, right? Um, what types of accounting work would you say are the easiest places for a company to to have that agent start to own it? And that that would meet their risk-off or risk-averse positioning in general, but still allow them to start to get a taste of oh wow, this is working.

SPEAKER_01 18:23

Yeah. I mean, look, I had a 15-year finance career, so me trusting AI and hallucinations, and you know, does my auditor okay with us if you're audited? Like, I get all of that. I hear you loud and clear as a fellow finance peer. Um the the the actual easiest place to start is just for non-finance tasks. So like have it write an email or you're giving someone performance review of feedback, have it synthesize, you know, all of the feedback into summary themes for you. Like you can just start there. I think within finance, the easiest place to start is have it review your work. Um, we have found like, you know, even if you're not in Campfire where it's built in, you can drop here's the 500 transactions I've coded. You know, do you find any inaccuracies? And it'll be like here's the three that that are off. And so it's like you can go the other way around. You don't you don't need to start where like it's just I think everyone thinks of AI and they think of like it's running autonomously on my data and it's like auto doing work, and it's like the biggest nightmare of like, oh, it's gonna create a huge mess in my accounting data. We don't need to start there at all. And then it's like once you have it reviewing your work, okay, we're gonna introduce the first time it's gonna do work for me. Then, you know, and within Campfire and with most systems, you can put controls on it where like every single task it does, it sends it to you for approval. So then, like, as opposed to you doing the work yourself, you get a promotion and you're managing an agent. It's like you have someone on your team and you're in a review state, not a data entry state. And so then you're just like checking its work and you're clicking approve every time. And then once you've clicked approve a hundred times in a row where it's been accurate, okay, let's move to like if it's over a certain dollar threshold, I will review. If it's below a certain dollar threshold, then I will auto-approve. And then we can continue on the journey. I think the biggest challenge I have with internal employees and with customers, when they send one prompt and super short prompt, like show me cash flow, and they hit send, and it's like it goes back to 10 years of data, and it's trying to do cash flow for like because they didn't give it a period of time, they didn't tell it what they want it to look like, and they're like, oh, this is wrong. And then they're like, AI is not ready for me. If you hired somebody and said, show me cash flow, and you you wouldn't fire them on the on day one. You would be like, here's what our here's what cash reporting looks like at the company, here's how we like to work, here's what, and then you would review the work, you'd give them feedback because odds of it being right the first time are wrong. You would train them, and then by month three, you're probably barely reviewing their work, hopefully. And so I I think about AI is the just treat it like a new hire on the team, learn how to work with it, learn where it's good, don't trust it in the beginning, but then learn over time of once you see consistent output, because otherwise you're not getting value. Like you ideally, you want the you want people on your team, you can go on PTO and like everybody still operates, you know, on their own. Like that's as a founder, like that's kind of what I hope for. Is like I can go on PTO and nothing breaks at Campfire. And so I think it's like, how do we get AI to be more autonomous where I can trust it, where I can enable it? Um, but also, of course, maintain accuracy of output. And that's through like a lot of iteration. So like I spend a lot of nights and weekends when I do most of my work with AI, tons, I give it tons of feedback. Like this morning, I probably gave it 30 areas of feedback before I got the right output. But now I I'm happy with the output. Now I moved it into a dashboard. And that dashboard, I now have an auto-refreshing daily. And now I don't, you know, maybe I'll spot check it or I'll like sniff, you know, if I see something that looks off, I'll kind of dig in. But ultimately it's like we gotta, we gotta really work with it and keep going.

SPEAKER_02 22:28

So for the listener, the the information that John just gave, like, love it. Start with your own work, especially if I mean, regardless if you're finance or not, like I think that's fantastic advice. Who's gonna know whether or not AI is doing a good job on your work better than you? So great advice. And you know, I think we've all heard the thing, treat it like an intern. But the way that you positioned it is would you fire that person on day one if you gave them crap instructions and they gave you crap output? Absolutely not. So for those of you that are going to start using it on your work, expect that the output that you get initially isn't going to be the best.

SPEAKER_00 23:06

Correct.

SPEAKER_02 23:07

You every time and and what what John is saying here is like he spends a lot of time helping helping the AI understand, calibrating it. No, actually do it this way. Oh, what I mean is this like all of these things, this is part of fine tuning generative AI use in your own role. I think that was fantastic. Advice. Now that we talked about what are some of the places where you know getting started and that sort of thing, but what are some areas within finance and maybe at large that you think humans still need to stay firmly in the loop?

SPEAKER_01 23:40

Certainly in terms of like synthesizing data from the AI and presenting it to other humans. I think one thing, you know, I just candidly struggle with was once AI really took off within Campfire, the output of, I mean, everybody went from a one-page memo to an 84-page memo. And now I'm consuming like massive quantities of content that I'm then like refeeding back into AI to just like shorten it back to one page for me. And so I think humans really need to be like truly accountable for the work. But also like someone did that the other day, and I was like, page 38, like I don't understand what this is. And they're like, oh, I didn't review it. So I think human need humans need to be accountable for the output of their AI. Like a member of your team, when I present, like our board meeting this week, when I presented the slides of the board, AI heavily helped in it. But when they're like, Where did you like what does this number mean? And I say, I don't know, like that's obviously unacceptable. And so I think owning the accountability, going back to the human on your team example we just had, own the output and be accountable for the output and be able to speak to the output as if it's your own. Otherwise, like AI slop is introduced into the system, and then you personally kind of lose trust of your stakeholders. So I think how you present it and and own it and review every detail continues to be critical. I think in terms of like taking those insights, like I mentioned earlier, like the five insights that you know Campfire produced for me to go action on. Like I'm not having AI like action on them with anybody else. Like it's up to me to go sit down with our marketing leader and go like execute on on what came out of there. So I think you definitely need humans there. And then I think judgment calls. Like, you know, AI might actually make a great piece of advice, but it's missing context that you have. And so you, you know, it uh absent that information and made a correct sound decision, but you have a broader set of knowledge. And so your ability to like synthesize, take recommendations. Like, you know, I wrote uh an email and before I sent it, I sent it to AI, and AI didn't kind of AI had read the email thread, but there was context that maybe was missing. And so it's up to me to really kind of come in and and say, like, no, this is actually what we're gonna do differently. So ultimately, you know, you're s it still reports to you. And I think everything I just described falls under that theme.

SPEAKER_02 26:14

So for the listener, yes, this uh started out as a finance uh theme conversation and how technology AI has impacted finance. But just in general, John, the the approach that you're taking and and sharing how you're using generative AI in particular, I think everybody like regardless of your role, you should be paying attention to this. This is extremely sound approach for sure. Um now one more question on this environment. How should the approval process, because if the if the human's gonna approve it and and it's gonna be AI generated before they send it off to you to present to the board, that sort of thing, what should that approval process look like between AI and humans?

SPEAKER_01 26:52

Yeah, it really depends on the task. So if it's just like like uh accounting transaction like within Campfire, if you're audited, you might like need an uh actual like human to click approve. If you're like not audited, like maybe you can just autonomously have it go through if the accuracy is there. For for other tasks, I think the approval ultimately comes of like when I build something with AI, um I own and like I really dig in on the numbers, and so I'll go like sanity check it, what's called like like a non-AI way. So I'll go into the system, like I'll go into our CRM or I'll go into Campfire, you know, our ERP, and I'll just like manually pull the report. And I actually just like download the report manually and I feed it back into the Ember Agent or into Claude, and then I have it tie out all the data for me. So I'll be like, here's the board deck. Can you go through every number and tie out every slide against itself? But also here's the manually downloaded data. Can you actually also tie every slide out against this? And so then it's like able to help me with that rigor and review process. And I think that's for everything that you do. So if it's even like an employee review and it's like here's the five themes from six people, I'll go into our system and like, was there actually six people that left feedback, or did it miss one? I think the biggest problem I still have, there's still hallucinations. So I was in our CRM the other day, and I was asking it questions via Claude, and it was using the employee ID. And at the end, it was adding the employee name to the employee ID. And so it was using the employee ID to build the report, but then at the end it added their name. So at the very it was like employee one, two, three, four, and at the end it threw the word Chris over one, two, three, four. And what ended up happening was it switched the names on the IDs, and so the employee, the sales team's rankings were actually the wrong names. And I was all about to share kind of like compensation rankings, and it was all wrong. And so, but I I and so that that's obviously the scary part for for everybody, and that's the one that I still struggle with, and that's kind of what we've really focused on with Campfire of like accuracy of output. But there's still a lot of like like, oh, you know, Claude didn't paginate the data, and only 500 rows of data were in each kind of API call is what it's called, and it missed like but there were 584 in the data set, and so it missed 84 transactions. There's like a lot of like hallucinations that are maybe it's just missing info, or maybe it has all the right info and it's confusing itself, like the salesperson's name example that I gave. And so, like your ability to like also be the human and voice of reason of like this doesn't look right, or using other methods of checking the work, like I described, of like downloading and uploading. Um, you still have to put a lot of rigor through it. And we've built a lot of that in a campfire to solve it for our customers. But you know, when I'm in other systems, like there's still a lot of hallucinations, and so it's an area I spend a lot of time because God forbid you get something critical wrong, you know, then then AI can be incredibly destructive.

SPEAKER_02 30:15

So what is the reaction of these finance teams when they it's beyond the sales demo? They're actually using your service and it's working for them. Like, is there a typical uh thank you letter or or testimonial that you get from these finance executives?

SPEAKER_01 30:33

We're fortunate that our customers love us and I get a ton of just emails and Slack messages and uh LinkedIn messages. One the other day was a CFO and he was like, I was on vacation and a sales rep was very upset about their comp. They wanted, they they felt it was wrong. And just on Campfire, actually, we don't have a mobile app, but they were just on the road. They just asked our AI, like, can you confirm this data? And Ember wrote answer with charts and outputs and confirmed all the data. And they just like screenshotted it and sent it to the person over email. Um, and they sent me a thank you note, like, hey, I was just on the road in the car, and was able to give them a an and it found seasonality. Um, it was like a holiday that, and so their their numbers were down, and it was like a short month, and it like, oh, their top customer, like, here's what happened. And the person, of course, like verified the output and it was all right. And they and they wrote me a thinking note of like, wow, this would have been like, you know, a couple hours of research for me. Um, but I was able to just do it in the car on vacation. And so just like the sheer ease of data access and kind of democratization of data, but also just people are like, you know, in accounting, the fourth of July in America is like always a working day because like month end, quarter end for a lot of companies, fourth of July, we just had um, a lot of our customers said I was able to take the fourth of July for the first time in my career. Wow. Because with Campfire, I went from a 15-day to a three-day close. So we were closed by the fourth. And so I think these are the moments that like whether it's more PTO or it's like just like better insights for the team, or it's the treasury agent I mentioned on just like saving the company money. You know, you're a hero in the next all hands. If you're like, yeah, hey, we got we just found more money by just building an agent. And or the sales rep's like, hey, that was like within 10 minutes, I was freaking out and they gave me a great answer and they told me exactly what happened, and I didn't wait two weeks, like kind of letting it fester. I think we can just like help in the moment and add a lot of value. And uh finance goes less of being what's called like a cost center and a little more of like strategic and like value add for the organization.

SPEAKER_02 32:45

So this kind of opens up the uncomfortable but important question um of whether AI is augmenting these teams or replacing parts of them.

SPEAKER_01 32:57

I would say nobody like there's certainly a shortage of accountants and in the news, it's like widespread that like people are leaving the accounting profession and it's low pay relative to some of the other roles. Like new graduates are looking at there's a lot of years of training, off in the CPA into low pay, and then you're like manually labeling you know cash transactions for the first you know, often three to five years of your career. And um I've honestly I saw a report the other day that um more people are entering the accounting profession, like for the first time it's seeing an uptick because AI is helping with a lot of the like call it monotonous or like you know, the bean counting you know, yeah, bean counting. For I'll use your phrase there. Um and I think that's incredibly compelling to me of like if we're gonna really like breathe life into the profession, it's gonna be let's make it more strategic. Everybody ends up being a manager as opposed to like manually labeling cash all day. And folks spend more time with peers, like with sales or marketing and helping like drive the business. And theoretically, that should lead to like more interesting work because it's more strategic, but also higher pay because then this the CEO is looking at you and you're up for promotion and you labeled cash correctly, or you like saved money by like building treasury agents. I you know, I think the one getting promoted is the one that like I automated cash with AI and I spent a bunch of time like growing us faster, more profitably. It's like, wow, this person, you know, needs a race. Like this is stuff that no one else was gonna do here. And so I think like I think about the profession is shifting as opposed to like roles being eliminated, and that's been consistent. Um, but I I think the overall theme is like in the AI era, most finance teams just aren't getting a lot of headcount. No one's saying, like, oh, AI's here, let me throw a bunch of people at the accounting team. Sure. So the accountants are often just working longer hours with no headcount or help on the way. And so they come to us and say, I do want Fourth of July back, or I, you know, I miss you know nights with my kids. And and we can be helpful there and we can lead to more strategic work and it's more interesting. And and so as a fellow kind of finance um peer for my career, I think that's my mission.

SPEAKER_02 35:33

Love it. So so if this is kind of the direction that things are going, and it makes perfect sense. What are what skills do you think that an accounting and finance professional, since more are coming in, what skills do you think they should be building now that maybe they didn't have access to, didn't have need to, or weren't ready for when they were entering the the profession?

SPEAKER_01 35:53

Yeah, I would say being what we call like system forward or or even AI forward. Yeah. And it's folks that are like like really attuned to everything we've discussed, like how to use AI and with while maintaining accuracy, how to automate, call it like transactional accounting work, like labeling cash or like tying out, you know, or doing a reconciliation, like folks that are coming up and really just like, I'm gonna on day one, I'm here to build a bunch of agents and I'm here to really just like use the latest and greatest tooling. I'm not gonna buy legacy ERPs and go manually do a bunch of work. I'm gonna like automate it all and really focus on managing agents. And so I think the even entry-level work is moving more from specialization to generalization because you're just managing like a team of agents that are covering more ground. And so each human has like a broader patch, call it of the world for all functions. I think whether it's legal or marketing or engineering or or finance, just everybody has kind of a broader remit and expectations are going up. And so folks that come in to an interview, and people often tell me I got the job because I pitched campfire in the interview. I think people want to hear you say, here's the AI native stack, here's how I'm gonna change the function, and here's how I'm gonna bring change. And by going from a fifth, because if you go from a 15-day monthly close to a three-day close, 12 days out of call it 22, 25 working days in a month, you're getting a third of the year back. January through April, a third of the year. You're now focused on other tasks. I think someone's gonna say, wow, please, like you're hired. I think that I mean that that's what I would say. You're showing like real ROI. Sure.

SPEAKER_02 37:44

So for those listening who are interested in uh profession in finance or career in finance, do that the exact same thing. Hey, I've got access to AI tools that are able to do boom, boom, boom, boom, boom. They're gonna go that guy. Nice work. This is like impressive stuff, man. And I know that that you guys are you don't need me to tell you that. Let me put it that way. You guys are uh on fire out there with some of the clients that you're working with and the the support that you've got from um really savvy investors and that sort of thing. So thank you for taking the time to um help me better understand what's going on over there at Campfire. But also, like I said, you touched on things that have nothing to do with your role as a founder or anything like that, but just like as a savvy user of AI and knowledge work. So um definitely some takeaways there. And for the listeners who are maybe they're wanting to run to their CFO and say, dude, have you heard this yet? What should be a next step if they wanted to explore more about what's possible with Campfire as compared to what they're currently using?

SPEAKER_01 38:43

Yeah, thanks for asking. Look, we we would love to chat. So check us out at campfire.ai and you can book a demo in the top right, but we've got a a demo on there. There's a there's a guided one. You can look at um other companies on the platform. We have all industries, we're live in every continent globally today. And uh, we would love to chat. We can get you in a sandbox if you want to check it out and show you and hop on and uh call with us and tell us your most hated task or tell us the thing you want to automate, or just walk us through all of it. It often turns into a bit of group therapy. Uh sorry, not group therapy, it just turns into therapy. Um, and we'll just show you the power of AI for your day-to-day workflows and things you haven't even thought of. So I love it. I'd love to chat.

SPEAKER_02 39:25

It's great, John, and we'll make sure that that's in the show notes for sure. If I was a finance professional, what I just heard would have me very interested in getting on that call. So again, man, thanks for taking the time. I know how busy you guys are and how fast you're growing and all that jazz. And um, I know that our audience is going to be uh glad that they learned more about what's possible just in general, but specifically about campfire. So thanks, man. I appreciate it.

SPEAKER_01 39:46

Thanks so much for having me, Chris.

SPEAKER_02 39:48

And for everybody listening, uh, what can you do to support us? Let somebody know that you heard this episode. Pass this on to your the finance team and your own company, or if you have some uh at the 19th hole next time, uh, if you have any finance friends, let them know that this is out here. Um, I think that the biggest part of this learning curve for all of us is sharing what we're learning. Don't assume that everybody, oh, everybody knows that. Trust me, they don't. Stuff that John shared today, like perfect example. Just different perspectives on on how to approach this stuff so that we can all get the weekends back with the kids, not spend 12 days on the month and close. Get the first fourth of July off in a long time. So um fantastic stuff. Thanks everybody, and uh, we'll see you on the next episode of Using AI at Work. Thanks for tuning in to Using AI at Work. Don't forget to subscribe for more conversations about how to use AI at work. And a special thank you to our sponsor, Chief AI Officer, for empowering businesses with AI education and training. Visit their website for a free AI readiness assessment and AI strategy guide to help you get started using AI at work. That's www.chiefaiofficer.com. Follow us on Twitter at the handle UsingAI at work and visit www.usingai at work.com for free resources to help you harness AI in your role.