Table of contents Toc Icon
Table of contents Toc Icon
In this insightful session hosted by Ace Cloud Hosting, Seth Fineberg, founder of Accountants Forward, speaks with Jan Haugo, founder of Smart Accountant AI, about how accounting professionals can adopt AI in a practical, firm-ready way.
With more than 20 years of experience running an accounting firm, Jan brings firsthand insight into the operational challenges firms face and how custom AI workflows, internal knowledge systems, and automation tools can help solve them.
During the conversation, Jan and Seth explore practical AI use cases for accounting and bookkeeping teams, including sales intelligence, workflow management, research, advisory support, and team training.
They also discuss the importance of secure cloud workspaces, responsible AI adoption, and how firms can build systems that enhance human expertise rather than replace it
Transcript
Seth Fineberg: Good afternoon. Or good morning, depending on where you are all dialing in from today.
If you’re here, you are in for a treat. Because not only are we going to be talking about AI, we are talking about AI for you, for accountants. And just kind of removing a lot of the hype, a lot of the talk that we hear around, you know, AI and why it’s great for the profession.
But we are going to see how it actually works from a practitioner herself who actually, you know, built some tools specifically for bookkeepers and accounting professionals alike. Cannot wait to get into that.
I want to first thank our hosts today, Ace Cloud Hosting. A couple of housekeeping items to go over. First of all, hi, I’m Seth Fineberg. I am working with the fine folks at Ace Cloud Hosting. I have my own business, Accountants Forward, in which I do content marketing and PR and, overall, you know, accounting profession advocacy to help see the profession progress.
So, today, while this is not for CPE, we just want to make that very clear. We do appreciate your time; we are respectful of that. We have blocked out a full hour for our host’s presentation today. But, you know, again, we are gonna have some poll questions just to kind of level set with you all, but this is not, you know, for CPE; this is rather really for your information, for your use. And I think you’re gonna get a lot out of it.
So, what are we covering today? Well, as you can see here, we’re going to move on to talk about how you can move from AI curiosity- maybe you’re already kind of toying around with some things in your own practice- to practical implementation, which I mentioned before: a live demonstration of how AI works.
Custom AI bots, you know, built for accounting profession use; how to approach AI adoption while maintaining data security- very important: accuracy, control, and professional judgment, and of course, why having a dedicated cloud workspace is essential when you’re building anything, something that Ace Cloud Hosting specializes in.
Our speaker today is someone I’ve known for a little while, whom I’ve come to know and respect. She is the founder of SmartAccountant.ai, Jan, and is our speaker today. Jan Haugo is, you know, really kind of leaned into AI in her own practice. As you can see here, she’s got more than 20 years’ experience, you know, running a firm, and understands firsthand a lot of the frustrations and questions that the profession tends to have around AI.
She also, you know, has worked with folks at Intuit on their AI certifications, and we’re going to be getting to Jan in just a moment. As I said, this is me. You know, I’ve worked in the profession, or seeing the profession as a journalist, for 3 decades, and so I’ve gotten to see a lot of the changes that the profession and technology impact in the profession has undergone.
We are at yet another inflection point in the profession’s history, with seeing how AI can work for us. And then, of course, our hosts, Ace Cloud Hosting, a leading managed cloud service provider, purpose-built for Accounting firms, are the folks that are behind everything that a firm is doing today, making sure that you are able to do it seamlessly and in a secure fashion.
So, we’re gonna have some offers coming up from Ace Cloud Hosting at the end of this. So I wanted to just get a quick read on everyone. We’re gonna get into our topic in a moment.
But I think it’s a good time to kind of launch our first poll, just for everyone who’s here, just to kind of level set with everyone’s thinking. So, if we can have that. Alright, here we are.
Poll number one! Where are you currently in your AI journey? It’d be great to kind of know, and we’re going to share results at the end of this. So, as we can see here, you haven’t really started yet. Find that a little hard to believe, but okay.
I’ve implemented the tools with ChatGPT or Claude, or others. I use AI for some individual tasks. I know that I certainly have, and folks that I’ve run into definitely are building things and using it in some very limited ways.
Actually building, you know, repeatable AI workflows that would be more advanced, and I’m already testing custom bots and agents in my firm. Gonna leave this open for a few more seconds. Pretty self-explanatory here. Just wanted to talk through those, as you all have a chance to weigh in. Very interested to see where everyone here is at with their AI journey.
What are the results? Of our survey. Okay, great. So, none of you are in the haven’t even started phase yet. Excellent. We’re leaning in. And this has really been a lot of my conversations with practitioners over the last year or two, where they’re at with AI.
So, as I said, at the end of this, we are going to have some very special offers for you to take advantage of, from Ace Cloud Hosting, both from a, you know, a hosting perspective and a managed security perspective. And also, $50 off an AI bootcamp. Our speaker today can talk a little bit more, and I think it’s time that we meet her today. Jan? Take it away.
Jan Haugo: Hello! Thank you so much; what a great intro. Yeah, I have known you. I was thinking about that when you said, “We’ve known each other for a little while, right?”
Seth Fineberg: Yay!
Jan Haugo: It’s probably been a good 15 years or so. But to be honest with you, it’s such an honor to be here. Yeah, give or take. It’s such an honor, though, to be here today, and I love talking about AI, so this is my favorite topic, and when Ace Cloud Hosting approached me with, Seth, I thought, this is a great opportunity to talk about kind of the basics, but also to show you some of the things that I’ve developed.
And so, I thought, you know, we’re gonna kind of move a little slow, and then at the end, we’ll have some questions and whatnot. I have been training in AI for the past two and a half years.
And what does that mean? It means that there wasn’t any formal training about two and a half years ago. I started studying and learning from outside of the industry, not inside of accounting.
And I’ve been building and growing 40 hours a week for 2 years, so that’s pretty intense, and I’ve seen a lot of changes, as you guys probably have as well, right? And for me, when Seth comes in and says, wow, we’ve got, you know, a bunch of people that are looking for AI, I always wonder, what are we looking for? So that poll is great.
Seth Fineberg: Yeah, no, it’s really great to see that, you know, just anecdotally, you know, we collectively will know where the profession on the whole is leaning in, in a way that we really didn’t see initially.
The first few years of cloud were like, I don’t know, you had a few people kind of checking it out and maybe putting some things there, but it took some time. But with AI, the level of control that practitioners are feeling, and everyday people really are feeling that, wow, I can kind of create something that maybe, you know, solves a pain point for me, or I’d be curious if it can do this.
And, you know, so you have a lot of the research side, and then you have what you’re gonna show us. So, I’m gonna stop, you know- can you share your screen now? Do I need to stop sharing?
Jan Haugo: Sure, yeah, if you want to stop sharing, yep, absolutely.
Seth Fineberg: I won’t get too. Good to see you. You’ve got the fun stuff.
Jan Haugo: I’ve got the fun stuff, yeah.
Jan Haugo: Let me go ahead and get this going for everybody. I think, you know, again, this is really about having a conversation, and there are a lot of conversations that are happening, so can you see that?
Seth Fineberg: That’s what’s needed right now.
Jan Haugo: Absolutely. Can you see that today’s Conversation slide?
Seth Fineberg: Absolutely.
Jan Haugo: Okay, so I’ve prepared a few slides for you guys, and then I’m going to go into a live demo, so we’ll kick out of here, but kind of the things we wanted to talk about today are: why are most firms, you know, where are they, and why are they stuck on the AI ladder?
The four tools that are predominantly doing the amount of work right now that people are seeing- what’s security about? You know, the IRS just came down with some guidelines, and I think it’s important to talk about how we keep things secure.
I’ll do a demo of one of the things that I’ve developed, which is you’re not supposed to have favorite children, but this is one of my favorite children, and it’s a sales operations manager.
And then I’ve got a bookkeeper buddy, and this is kind of a custom GPT that I built to help mentor staff, but it’s also a way to get people comfortable with AI. And then we’ll address some of the
Seth Fineberg: Training side of things, right? I mean
Jan Haugo: Well, yeah.
Seth Fineberg: That’s Brilliant.
Jan Haugo: Yeah, talk about the training, and then we’ll talk about what some real concerns are, and then where we’re gonna go next. So, if you’re ready, I’ll kick it off.
Seth Fineberg: They’re ready.
Jan Haugo: Alright, sir. Let’s do it.
Okay, so when I kind of look at where people are on their AI journey, I love that everybody here has experimented and tried AI, and you’ve got the different levels, right? But what I see a lot of is people are using it as kind of a skill producer.
And what that means is they come in and just as Seth said earlier, hey, what could I build? What is this one thing that I need? The other thing I see, and I’m guilty of this too, is that I use AI like Google. What can you tell me? Can you please review this profit and loss? Can you rewrite an email for me? I need to know the difference between the restaurant down the street and the one right across from my house.
Like, those are the ways that people have been kind of using AI, and it isn’t that they’re not getting into the higher levels of AI; they just really haven’t been taught how to move up that ladder. There’s a lot of, kind of, one-shots: do this, do this, but it’s not really piecing together. So most, most people I know are using it like a skill, like a Google, like a great add-on to your firm, right?
Seth Fineberg: Yeah, it’s search, it’s, you know, a little it’s search, and you know. Like, what can we do?
Jan Haugo: Yeah, yeah. And honestly, that curiosity is fabulous, but in order to start seeing those efficiencies that you hear about on, you know, LinkedIn and all the other places, you have to start putting some power behind it. And that’s where you develop the brain component, and that brain component is really the context.
And what I see mostly from people is not giving AI enough context. They just kind of assume that AI knows it, right? And it doesn’t. That’s what the humans are for. That’s what we know.
We know the relationship with our clients, so if I have a client who’s got two kids, and he’s worried about putting them through school, and we need to help him with a cash flow forecast to start putting those funds away, I know that, but the AI doesn’t know. They just see the company that he’s running and what the statistics are. So that brain component of it is the context that people don’t give AI enough of.
And then when we look at it, a lot of people have, in the very beginning, the connections weren’t great. But if you think about, like, Xero or QuickBooks, or Bill.com, we connect these things all the time. We connect to our bank, we connect to QuickBooks, we connect to Xero, we connect to our advisory dashboard tools.
So connections aren’t unfamiliar to us, but they are unfamiliar to us when we’re connecting them to AI. We don’t quite know what that connected space looks like. But not a lot of people are using those connections right now, and a safe place to start is to use a connection to your calendar.
And a calendar connection is great, because it can go through and mine the other day was said, you have a huge week coming up. I found 4 places where you can make some efficiencies and changes if you rework how your calendar looks coming up. And that’s something I never thought about before, but it’s a great, easy way to connect.
But when you connect that AI to your tools, you start using the brain, and then you implement the skills you’ve learned; you’re starting to see a lot more power of that AI coming through.
And then finally, the last one, which not a lot of people are really doing yet, is scheduling. It’s kind of creating that recurring schedule for that AI, whether it’s giving you a daily update, whether it’s giving you a calendar update, whether it’s giving you an overview of your emails. Those are nice, safe areas for you to kind of test with, but it’s also a great way to get the power of the AI.
And it sounds like, oh, great, I can connect to my calendar, and I can connect my email, but what people don’t understand is when you give it all four together, the power that you start to see with the AI will blow you out of the water; it’s insane.
Seth Fineberg: Are you finding that also, Jan, with Agentic AI? Are you using agents for anything? Or is this mainly Claude or GPT primarily?
Jan Haugo: Yeah. So, with the Agentic, if you look at, kind of, the three phases of AI, you’ve got the base one that we came out with, which was predictive, right? And why was that predictive?
Because we gave it all of the information that we could find, and we dumped it into this machine. And so, if you ever typed in, like, Google, and you said, what color is my? And you saw that list come up, it was, like, car, hat, shirt. That was kind of the first version of it.
And then we moved to generative AI, and that’s kind of where AI is predicting the most possible event that’s gonna happen next, and it doesn’t have a goal, it doesn’t have a lot of memory, but it starts to generate on itself.
So, the generative is that. It starts to learn a little bit from itself. And then when you look at the Agentic, this is where you add all three. It’s got the goal. It’s got the tools, and then it’s got the loop.
So you can have the tools where you connect it here, and you can have the brain, and you can schedule it, but in order to get it to do it over and over and over again, and to think and iterate on that, that’s called the loop, and that’s… the result of what we’re getting with this Agentic AI is it learns from itself, starts to build on itself, and it remembers that loop.
So, for me, pulling in data, connecting it, and having it build on that memory is super important. So, I’m not seeing a lot of people do that yet. I’m doing that. But it takes a little while to get there. The trust factor has to be there. I think the comfort you want.
Seth Fineberg: 100%.
Jan Haugo: Yeah, for sure. Because, you know, it works so fast, you don’t know what it’s done, and it’s done in the blink of an eye.
Seth Fineberg: Yeah, and, you know, we’re always dealing with, you know, it’s only as good as the data that you have. I mean, if somebody has a lot, how do they, you know, do you hear this, too? People are like, oh, I don’t know what I should put in. What shouldn’t I put in? That’s a pretty common concern.
Jan Haugo: Yeah, very common. A big concern, yeah, absolutely. Trying to understand what to put in, what information, you know, everyone hears about security and safety- you know, what do we put in, what do we not put in- so that’s, yeah, for sure, I hear that a ton.
Seth Fineberg: Yeah, I mean, how are you dealing with that with what you’ve built? I mean, were you able to just kind of sandbox, you know, all this, actually have a secure environment?
Jan Haugo: Yeah, let’s see, I think we’re- I’ve got a slide here that talks about it, and here we go. We designed this grid.
Seth Fineberg: Almost as if I knew what was coming next.
Jan Haugo: Almost. Yeah. So, I kind of put it in a red, yellow, and green perspective, right?
And so we start with the green, and this is always safe to put in the AI, so drafting emails that don’t have client-sensitive information. SOP documentation, concept explanations, like, you know, I have a fixed asset that needs to be depreciated, I’m not sure how long, you know, that type of thing. Template creation, creating templates over and over; sales and marketing data that have no PII (personally identifiable information).
And what that means is, if I have a name with a bank account number and a bank, I can put those all together, and I can identify that person, I can identify that bank account, I can identify the bank, so that’s PII.
So if you don’t have any PII in any of the information you’re using, then everything in that green zone is safe. In the yellow zone, this is where we start to redact information, because we start to hit identifiable information. So, if we put in a trial balance, I will run a report in QBO, and I’ll take off the name of the client.
I’ll also go through the chart of accounts. I have a lot of people who will put, like Chase, and then they’ll use the last 4 digits of the bank account. I’ll take it out and just say bank, you know, basic checking, basic savings.
So, you want to sanitize some of that information, and then you can put that into the AI. Same with the P&L. Remove the headers. I have seen, like, a balance sheet, and it’ll have, like, a loan to Joe Smith, you know; you want to take out Joe Smith’s name from the transaction list.
You can definitely put it in a general ledger, but you want to take off the company name, and then you want to scan that bank text, or that memo field; often a bank text will come through, and it might have the account number in there. So you want to kind of go through.
Seth Fineberg: Right, I’ve seen that.
Jan Haugo: Yeah, yeah. Categorization help, again, trying to, you know, kind of make sure, like, we’re not coding it to Joe Smith’s loan, that kind of thing. And then, if you’re not sure, then we treat it as red.
And the red is where we just don’t ever put these things in. You don’t ever put in a Social Security number, an EIN number, a tax ID number; never put in bank routing numbers, full credit cards, payroll details, obviously with employees’ names, or login credentials. So those are the levels, and if you’re ever unsure. Treat it as a red.
Seth Fineberg: Yeah, that’s automatic red. This seems like something to have on your wall, or on your desk, or wherever you’re already, you know, because it seems like everyone here is, you know, doing something with AI. Have this ready. Love it, Jan.
Jan Haugo: Yeah. So my thought is, if it would appear on a billboard, would it hurt someone? And if it’s yes, then don’t put it in. You know, if no, then you’re safe. Again, just trying to be secure. There’s not a lot of, you know, regulation right now on this.
And always, always, always use the paid version, because if you use an unpaid version, the $20 version, whether it’s ChatGPT or Claude, then they don’t use your data to train on, the other models, the large models, when I said that they dump all that information in, they don’t dump your information in, it doesn’t mean that your information isn’t out there, it just means that they don’t train on it, and so I want to make sure that that’s clear.
Seth Fineberg: Yeah. And what about, like, if AI can- and I know we have all the fun stuff to get to, but since we’re on this topic of safety, you know- can AI, you know, still leak data, even if they get an enterprise account with data training turned off? Yeah, you heard about that.
Jan Haugo: Yeah, when I was at Scaling New Heights, Randy Johnston, who we both know, did a great session, and he said, you know, he had put some information in to test his AI to see if it would come up later. And he said he put that information in, you know, and was secure with it. It wasn’t the red area, but it was something kind of maybe yellow and sanitized out. And it wasn’t accurate information. It was just something he made up. He said a few chats later.
Seth Fineberg: It could have been, yeah.
Jan Haugo: He said, you know, a few chats later, I don’t know, maybe a couple weeks later, that information resurfaced, not just in his chats, but other people’s, that it was kind of sitting out there as kind of a reference point or whatever. So, I definitely think there is a bleed of that information.
Seth Fineberg: So it can’t.
Jan Haugo: Yeah, I think it can. I haven’t proven it; Randy did. But yeah, for sure. And that’s why I kind of go back to using, like, Google and Microsoft, those enterprise versions, so you already have Copilot, you already have, like, Gemini. Those are enterprise versions, and as long as you’re using their tools within there, not pulling data out. Those are much safer. The information is still in a bubble. It’s more secure, but it’s not, you still be cautious, still use this kind of grid.
Seth Fineberg: Yeah. Alright. So, yeah, if we’re about to oh, here we go. This is, yeah, the next one.
Jan Haugo: This is the next one, yeah. So, people often ask me, like, what do you like? And I use them all. And, you know, I’m lucky enough to be in an environment where I get to use them all, but at the same time, they do have different, you know, uses. So, Claude, you’ll look at that; it’s very structured, it’s got a lot of multi-step reasoning, it holds a lot of context across an entire build.
So if I’m building something out, or I’m kind of iterating on a conversation, or I’m using, like, my sales operations manager, I’m going to use Claude. And you’ll hear people say often, it sounds more human. You know, you’ve used it, right, Seth? So
Seth Fineberg: Hell yeah!
Jan Haugo: Yeah, yeah, it sounds good. I think for writing, you know, for reasoning, I think it’s beautiful. If I am looking at fast conversations, I’m looking at, I love ChatGPT for Excel; great, like, their Codex is amazing. Trying to get in and have it do a lot of Excel work.
A lot of critical thinking. I love ChatGPT for that, but I also love it because it’s very shareable. So in an environment where I want to share it.
Seth Fineberg: Within, like, your team, yeah, and your clients, yeah.
Jan Haugo: Or even my clients, yeah. We created something called a Bookkeeper Buddy, which I’m going to show here in just a second, but a lot of my clients don’t have an AI policy, right?
So, in my bootcamp on Monday, we created an AI policy generator, so you can use it in your firm, but the great thing about ChatGPT is you can share that outside of your organization, so you can share that GPT out to your clients and say, hey, I think you need to create an AI policy, right?
And then your clients can use AI to generate something like that. So, I love ChatGPT for that. Gemini is amazing, and I love it because you’re in the Gemini workspace, and it pulls straight from your documents and your sheets and your drive, and you just ask Gemini, and it goes.
Seth Fineberg: Everything you’re already working in, in that environment.
Jan Haugo: You already paid for it, too. You know, so you’re not adding an additional cost. And then Copilot, Copilot and I have this love-hate relationship, but this is Microsoft’s tool.
Seth Fineberg: I’ve heard that.
Jan Haugo: Yeah, but if you’re in a Microsoft environment, this is the guy, this is the one that you’re using, and I have to say that they are getting better week after week, sometimes day after day, but it’s the same as Gemini. It can pull from your SharePoint; you can go use it in Excel; you can use it in Word, and it’s amazing.
And again, the enterprise versions, not the personal versions, are the ones that have more security, so when you look at Gemini and Microsoft, you want to look at the business enterprise versions.
Seth Fineberg: Right, and I like that, you know, in just a tip of the hat to our host today, the answer, as you mentioned before, about, oh, is the paid version safer?
And this is why being in a secure hosted environment like the one at Ace Cloud Hosting matters, especially when we’re just kind of starting to lean in here, and you do have those security concerns. Again, go back to Jan’s chart that she created, the red, yellow, green.
Jan Haugo: Yep.
Seth Fineberg: Learn it. Know it.
Jan Haugo: Yeah.
Seth Fineberg: Have it on your desk, or nearby.
Jan Haugo: That might be my new tattoo. So, yeah.
Seth Fineberg: If you do that, I would love to see it, Jan. So, I think we’re about to get into your demo, so I wanted to just hit one more. We’ve got another poll question coming up right before we jump into that, Poorvi, if you would. Share that.
We’ve got poll question number two. Yes. Where would AI create the most immediate value in your work? Jan’s gonna share with us what she created in just a moment. But for you, just as we’re starting to think about these things, is it with repetitive accounting processes? There are a lot of them.
Client communication? Doc review, information extraction, research, and summarization seem to be where a lot of people are leaning in. And, you know, also internal productivity and workflow management- or sorry, we don’t have all of the above- but, you know, whatever one of these you would say stands out most for you.
Where do you think AI would create the most immediate value for your work? And then we’ll get to what Jan created. Leave this open for a couple more seconds. So, we get everyone signing in today a chance to answer. Today’s poll.
Okay. Let’s close it up and see what we got see what everyone said today. Where would AI create the most immediate value in your work? Alright, good variety here. We have some repetitive tasks, a bunch of folks saying research and summarization; that’s kind of what I thought.
I mean, this is, you know, stuff you need to know; you need to know it right now, and also a good amount of internal productivity and workflow management. Nice, thank you so much for participating today.
I will not keep you any further from our first demo, the sales department manager that Jan has talked about, and I personally can’t wait to see it, because I haven’t actually seen this, but Jan, this is going to be live, folks.
Jan Haugo: Yeah, if you ever come to anything on…
Seth Fineberg: No, no net, no net here.
Jan Haugo: I know, I’m like, Seth, balance me. I mean, you know, Seth and I talked about this. He’s like, we’re gonna do this live. I’m like, that’s the only way I know how to do this.
Seth Fineberg: Nice. Yeah, like, no slides here, people, like, just this slide.
Jan Haugo: We’re gonna get right into it. So, when I started studying how AI was going to change our lives, the immediate thing that happens first, that everybody looks at, which was on the poll, is replacing that repetitive work, the production side, right? Everybody’s looking at how can it enter my bills, you know, how can it kind of save me time in reconciliations? And to be quite honest with you.
There are a lot of companies that are solving for that. Intuit, Xero, Digits, Kick, and all of those places are working on that with their own AI. So why reinvent the wheel? My thought then was, if we’re reinventing the way that we do production.
That means, kind of, our back end needs to change. Our personal company, our accounting firm’s operations have to change. And you can’t just have a front end that’s efficient and a back end that’s still working on paper or Excel sheets or moving stuff around.
Seth Fineberg: Yeah, it’s not gonna work.
Jan Haugo: No, no, it’s not gonna work. So I built out 21 different components of my operations. I have a brand voice, which is my company’s voice throughout my entire set of operations, so it knows me, it knows who we are. I have a sales manager, a sales department, I have an HR department, I have a customer success team, I have a CFO, my own company CFO.
So we have all of these departments built out, and I’ve revamped them all, and they’re all in Claude. Claude, at the time, was way better than ChatGPT in terms of kind of running these operations systems, and, for my purposes, this is where I put it. You can put it into ChatGPT; I just don’t think it’s as strong for some of them.
So what I’m gonna show you today is my sales department manager. When I get a sales call now, we’ll go in and do some prep, which is something I was never able to do before when I was trying to do sales. I couldn’t really look up the person or the company that was talking to me.
Sure, you do a cursory LinkedIn look, or maybe you go to their website or whatever, but, you know, how do you really know much about the company? So, that was one of the things that rolled into my sales department.
The other thing is that I wanted it to kind of tell me what it thought were the points and the questions that the customer might have, and any kind of hook that I might be able to kind of say, hey, this is an offering that aligns with you. So, I used to spend a lot of time trying to get information on the prospect, and so now I just drop in my information and let it go.
So, I did a masterclass last week, and I did it on our sales manager. And I created a fake company, a fake CPA firm called Highline CPA. So I’m going to show you a little bit of that demo as well, and then I’m going to show you the other prospecting. So, I use my sales department here. Let me just minimize my slides, and I will pull it over for you.
And I use this guy, no other. So here we go. Alright, so I am in Claude, and if you can see, I have a project set up. So I have all of these projects; I have just a ton. So, like, if I look, here is my leadership and development coach; my sales guy is number 2, so we’ve got all of these projects that they’ve set up, and this is where that context and that memory that I talked about earlier comes in.
This is a great way to do this, but when you look at my sales wizard behind the scenes, you look at my instructions, and my instructions are detailed, so it took me at least, I want to say, close to 40 hours to develop this.
And that is a long time, but the benefit of what I save on the front end is amazing. So I tell it what the role is, I tell it what its key objectives are, I give it my decision logic, I tell it my timing and capacity, and how we learn now. Don’t forget, I’m taking this from the standpoint of I’m an AI trainer now, but I give it budget signals and pain point dominance. So, as you can see, this is quite a big prompt.
I wouldn’t expect anybody’s to be this long, but I wanted to include everything that involves sales training. What are my sales methodologies? What’s important? When people are thinking about urgency or scarcity, how do we talk? And what are my final notes? So, the key here is giving it that context, right? So that’s the first thing. You have to give it the context.
The next thing that I gave it was our knowledge base. And this knowledge base, it knows our pricing, it knows our writing formats, it knows our branded marketing, it knows what our master classes look like, it knows my brand voice.
I use Daniel Pink’s modern sales framework, so I roll that in; I use Chris Voss’s inspired techniques for sales. I have a belief-shifting framework in here. I took my bootcamp sales page, and it’s in here. It also knows my sales goals and our Black Friday specials. How much of this does your salesperson know?
Seth Fineberg: Yeah.
Jan Haugo: Right?
Seth Fineberg: Just for clarification for everyone, I doubt Jan just sat for 40 hours building. This was just something she did over time, but something you touched on, Jan, that I didn’t mean to interrupt, but I wanted to just reiterate that everything I’ve heard about building in AI and creating is as specific, and accountants, we know you can do this, as specific in your verbiage as you can be.
Jan Haugo: Yes.
Seth Fineberg: You know, if you think you can go deeper, ask it even, even deeper. I’m seeing some questions flow in now that we’re like, but what about this? But what about this? That’s the right mentality and the right approach to have.
Jan Haugo: Yep. Absolutely, and that inquisitiveness is what led me to put more files in. I don’t want to just overload it with information; you want to give it the right information. And again, I don’t expect anybody’s instructions to look like this, but you need to start with instructions. It needs to know who you are. And it needs, you know, like a trainee, if I’m going to bring someone in on my firm.
And I’m going to train them. What tools do they need to be successful? You’re going to give them those tools, right? So, I’m showing you the behind-the-scenes so that you get an understanding that this isn’t just a prompt, right? This is a true, built engine. That is the core of my sales component.
So, when someone comes to me, and they say, “Hey, I want to talk to you about AI”, the first thing I do- and this is a friend’s firm- I will take their website just like this, I will hit Ctrl+A, and I will hit Ctrl+C. Ctrl+C copies it.
And I will go in, and I will put this in here and say, I have a prospect that would like to talk to me about AI. Here’s their website; take a look and tell me how I can help them. Give me the good, the bad, and the ugly.
And so I paste that in. And I let it go to town. And so while it does that, what it does is it goes out to the website. What’s always amazed me about this process is that it finds things I would have never thought. So it says, look, they’re a real firm with actual scale and range, full CPA, different locations. Bad.
The site is stuck in a pre-AI positioning. Every claim is about people, our team, our CPAs. We treat every account as if it’s our own. There’s zero mention of technology, automation, or AI anywhere on this page.
Right, so what am I trying to work with them on? I’m trying to work with them on AI, so this is a great opportunity for me. The ugly. Two offices means two sets of habits. If it’s in Illinois and Arizona, they’ve grown somewhat independently, you’re probably looking at inconsistent SOPs across locations.
Exactly the kind of tribal knowledge, not documented process problem that AI can help them with. How fast did it take me to pull up this information? Right? And it gives me my toxins.
Seth Fineberg: Seconds.
Jan Haugo: Yeah.
Seth Fineberg: It gives you a kind of a snapshot. I mean, you can use personal human experience judgment on: Did it get this right? Did it leave anything out? Is this enough for me to have a conversation with? Yeah, somebody was very, very curious about the output here. Can we give one more example?
Jan Haugo: Yes, so, one of the other ones, and again, I want you guys to see it will tell me, do you want me to put a pre-call intelligence brief together? So, a couple ways before I move on are, number one, I can have it build out that pre-prep call. When I’m in the middle of the call with them, I can come in, and I put in objections, and I know the objections because I know what everybody is telling me, right?
I’m afraid of AI. My team won’t adopt it. My clients don’t like it. It’s hard for us to learn. We have to change everything. I can type all of that in real-time while I’m talking to the client, and this will help me with those objections.
So, this is like having a mentor sitting right with me, right? Amazing. Just amazing. Now, we did AI- the AI Highline accounting firm was one that I created last week. And, basically, I told them, we’re doing a demo for a Highline CPA firm. This one takes a little bit longer, but I want to say you’re a growth strategist.
I want to help them analyze competitive analysis. So, one of the things I never really did before when I was in my accounting days was I would go around and look at the competition. Every once in a while, I would, but not often. So, I will come into my sales guy.
Seth Fineberg: Good sense of, yeah.
Jan Haugo: Yeah, what are other people doing? Yeah, and so I will come in. I’ve already searched for accounting firms in this area, and it will ask me to go in and do a competitive analysis. How is the guy down the street working on things for me? What are they offering that I need to offer? What’s going to help me kind of move the needle? Who should I target?
Because he’s, you know, 10 miles down the road, right? So, we’ve got some logistics, even though most of us work virtual now. There’s a lot of opportunity, so I’m gonna just say we want to do client-facing, persuasive- let’s see- full competitor pricing table.
So, I will go out and search all of my competitors; give me pricing across the board. If it can find it, it will tell me what their offerings are compared to my offerings and allow me to start to build out a program, a product. Help me with some of the marketing for people that are close to me.
Now, there’s a ton of work to go around, but you want the right customers. You just don’t want a customer.
Seth Fineberg: Right.
Jan Haugo: So, how do we get the right customer? We offer them the right products. And this guy, my sales guy, goes and does that legwork.
And I had someone on the call with me on Thursday, just an attendee like you guys, and she did competitive research for a large Fortune 500 company. And I asked her, I said, how long would it take you to do competitive research? And she said, Jan, it takes me 3 full 8-hour days to do competitive research.
And it’s like, wow, that’s a lot of work, right? And in one minute, this finishes. And in one minute, I have my competitive research that took her 3 full 8-hour days. So, how can you not use a tool on your back end that you’ve built that knows your company, that can help you drive growth, can help you drive sales, can help if you bring in a new salesperson, can mentor your team? How can you not use a tool like this and gain a leg up? It’s pretty easy.
Seth Fineberg: Excellent. I know that it does, and I am excited to see this, as well as your bookkeeping buddy.
Jan Haugo: Yep.
Jan Haugo: While this one’s working, we can work at Bookkeeping Buddy if that works for you.
Seth Fineberg: Yeah, I’m just looking at our time. I want folks to see that, and then also, you know, get in any other questions. I’ve been working them in just as we’re talking, you know, some of the questions that do come up. But, you know, if there’s any at the end, I want to get to those. All right.
Jan Haugo: Okay, so let’s go ahead and let’s look at Bookkeeper Buddy. So it did come through, and it just said, I’m not going to open it because it just put it in Markdown for me.
But essentially, I think somewhere along the line, I’ve got a problem with my network’s a little weak. But let me see if I can pull one up. Competitive analysis, the one that I had done for you guys before. Yeah, see, it’s not downloading my file right now. Perfect opportunity for a live demo not to work. But you get the gist, and here is the competitive analysis, right?
Seth Fineberg: You know, you’re at the mercy of your connections and things like this.
Jan Haugo: Well, and of the AI, but it does find me some pricing; it does tell me that they’re doing advisory tier work; it is letting me know what the customer framing is looking like, so where do I want to lead in? And honestly, if I could click on this, it does kind of show you the gaps and things where you could probably take an opportunity. This is kind of show you the gaps and things where you could probably take an opportunity.
But again, this is, you know, sometimes it’s working, sometimes it’s not for me, so every once in a while when I get on this.
Seth Fineberg: Yeah.
Jan Haugo: Even still, you can see that I’ve used it for a lot of competitive positioning, Sales Wizard overview, to review a lot of other things, to do a lot of different kinds of, you know, aids in growth, so I use it often. Yeah.
Alright, so let’s flip the script now. This is part of thinking about your back operations, giving it context, giving it the knowledge base. The next one I built was when our team, who was a little bit younger, was struggling with AI. They were struggling with how to get involved, how to use it, and how it was going to help me.
Seth Fineberg: What makes the most sense, where, you know, do I start? Do I start with Claude? Do I start with ChatGPT? Do I go right over to, you know, Gemini or Copilot, which, again, are very much like, well, what are you working in? Are you working in Google? Are you working in Microsoft tools?
Like, those that will help dictate, but the other ones, which you’re also paying for every month, hopefully, you’re getting, you know, the better versions that you are paying monthly for. You want to know what it can do. So, your bookkeeping buddy, you mentioned, was built using GPT.
Jan Haugo: Yeah, so I use ChatGPT on this one, and the reason I use it, again, we’re going to go back to that sharing was because I could share it with my team, and at the time, it was just really easy to share. So, behind the scenes on this one is very much like my sales agent. I have a huge set of instructions.
But again, you don’t have to start with each set of instructions. In fact, the team at Ace Cloud Hosting and I put together some ideas for you, and some prompts for some of these, but you can see that my Bookkeeper Buddy is built on progressive development.
The biggest thing that I’m proud of is that I built this out telling you, you’re not creating; you’re not just answering these questions; you are creating the next generation of trusted financial advisors who can think strategically, communicate effectively, and serve their clients with wisdom and integrity, right?
So, you can put anything in here, but I did not want this to be just, give me the answers. When I built it, don’t forget it was junior staff. But even still, I will make a journal entry backward 9 times out of 10, and look at it, the financials, and go, crap, I put my debit and credit in the wrong place.
So, I put in accounting basics, I put in adjusting entries, setting up zero, creating SOPs, I put in a bookkeeping cheat sheet, evaluating business, financial statements, improving a profits cheat sheet, payroll accounting. Are you guys seeing the thought process here that we give it the sources it needs?
We give it the instructions it needs. I’m showing you two different things, but at the same time, they both have a common theme. So, how we use this is I give this out to the team, and they come in, and we can tell them, you know. Any kind of question, give me the journal entry to book a $10,000 car.
And the cash came from the bank. Now it shouldn’t just answer; it should ask me some questions, right? But it did; it just answered. Tell me the why. It’s going to tell me the debit to the vehicle asset because the business acquired a car, the credit to cash, right?
So it’s going to come out, but if I want to say, I need to figure out how to develop a fixed asset or depreciation schedule so that I can show my work for my team, but also in the books. How did I do that, and can you help me develop that?
Right, so it already says, record the depreciation. I can ask it literally any accounting basic question, and it will come in, and it will act as a mentor and solve that question for me. So, it’s starting to tell me what the fields look like: my depreciation, my purchase.
My team loved this, because essentially, what happened is, they were asking the seniors a lot of questions, and these are things that the seniors felt that the juniors should know. While this was a great way for the juniors to come in. Do some of their own research.
Seth Fineberg: And not have to bother the seniors.
Jan Haugo: Confident.
Jan Haugo: Absolutely.
Seth Fineberg: It’s just, look, you get, especially when we all get focused on our work. You want to help out your junior team member, but
Jan Haugo: Yeah.
Seth Fineberg: Sometimes they might be waiting outside your office, or waiting until you’re free; maybe they’ll send a Slack on it, something, you know, depending how you’re communicating. This is, like, nope, it’s… Go ahead, you can get the answer here.
Jan Haugo: Yeah, no, I mean, it worked it all out. It gave us our depreciation schedule, our asset register. You know, you can ask it to create an Excel sheet, so you can back up your work papers.
But what better way for the team to learn AI than to use a tool that gives them confidence and helps empower them to be more independent, as well as giving them, you know, that boost into AI, right? So, for me, this was a huge win, because I could just share this out, and my team members could use it. I didn’t have to.
Seth Fineberg: How you tie it into the books, it’s right there. Wow.
Jan Haugo: Exactly, yeah. So, the other thing that Ace Cloud Hosting and I did was we put together an idea list of something like this that you could create. One of these that I created, which was really popular, was a W-9 and a 1099 questionnaire. So, we put in the resources from the IRS.
So, if people weren’t sure about how or who should get a W-9, and what box or which 1099 gets issued, or whatever, then you can come in here and ask it. Now, it doesn’t, you know, qualify as legitimate, you know; you still have to check the answers, is what I’m saying. You can’t just say, oh, this chat told me.
You have to check and verify it, but the thing is that it gives you that starting point. It says, hey, this guy made more than $600, you know, he’s an independent contractor, we need to get a 1099, we need to get a W-9 from him. So, that was a really popular tool. I’ve done one for advisory mentor, so a lot of advisory needs to start happening, but a lot of people don’t know how to do advisory.
So, it’s the methodology behind advisory. So, you come in and say, I’ve got a call with a client, I need to do advisory, how do I go about the conversation? And it interviews you, and preps you, and talks to you. Those are some great ideas, but it’s a great way to get over that hump of, write me an email.
Seth Fineberg: Excellent. Well, I know that we’re somewhat limited on time here. We’ve got a couple of questions I want to get to. I do want to hear Final Poll. I do want to hit our final polling question. Yes. And then, I want to get to some of the live ones that we have received, and they’re pretty good.
So, if we want to just do that, we’ll get to our final polling question of the day. So, given the sensitive data that your firm handles, where do you currently run AI workflows, custom bots, and agentic AI tools?
Local desktop or laptop in a dedicated virtual cloud environment? Not really sure. Good question. Since a lot of this, you know, again, you know, we’ve touched on the security aspects of what we do. Some is in its own environment already.
Jan Haugo: Yeah.
Seth Fineberg: But, you know, when you’re building things, you know, given the sensitive data that your firm handles, you know, where are you currently running AI workflows, custom bots, Agentic AI tools? So it’s either on your local desktop or laptop, in a dedicated cloud environment, or you’re not really sure.
Alright. Let’s see what everyone said, and we’ve got a few minutes for questions here, which is great. Alright, we’re going to share the results in just a moment. See where everyone’s at. Okay, kind of split between local desktop/laptop and dedicated virtual cloud environment; a couple of people, not really sure. You have been, you know, using AI in your work. Has AI replaced any part of your workflow, or is it mostly speeding things up?
That’s a really good question. This narrative of replacement is a constant specter in our profession right now, but realistically, it’s like, well, what’s it taking away? Or is it just speeding things up?
Jan Haugo: Yeah, I think it speeds the work up. You can work faster. I feel like people should actually take that capacity that they have gained and start learning, so allowing their brains to breathe. I’m not gonna lie, it’s a lot of heavy brain work. It doesn’t do a complete workflow.
I mean, I could have it do a complete workflow, but you know, there are some things, as my sales wizard comes in, and I get an email, and then, you know, it populates through. So, there are some components that we have automated, and honestly, for some clients, I’ve automated entire systems, but there are workflows that we’re automating. It’s not the entire company.
Seth Fineberg: Right. And I have seen where, you know, some of the Agentic platforms that are out there, they’re addressing that full workflow, you know, issue. Alright, I want to share just the rest we have here before we go. Okay. So again, I didn’t want anyone to really forget about some of the offers that we do have. And we still have a couple of minutes left if anyone has any other questions for Jan.
So, you know, please take advantage of some of these offers, particularly the AI Bootcamp. $50 off. A master class, was it, 90 minutes or so? Something?
Jan Haugo: Yes, 90 minutes. We do it every third Thursday of the month. It’s completely free. It’s just a way for me to show you, like, the sales wizard, or other tools that I’ve built, and kind of talk about, like, security and things like that.
So, we do the boot camp usually following that, almost immediately, sometimes the next week right after.
Seth Fineberg: Right. Yeah, we’ve got some other great discounts here, as well. A couple more questions, sneaking in here. Any use cases for tax pros, in particular?
Jan Haugo: Yeah, you know, tax is such a sensitive area.
Seth Fineberg: It is, but, you know, workflow is one of the things I hear come up quite a bit. And also, on the research side.
Jan Haugo: Yep, 100%. The research on these things, as you saw, my little sales engine just did a quick, you know, dive on competitive, but Perplexity, which is a huge research agent, you put Claude or ChatGPT into that deep thinking mode, and their research is outstanding.
So, if you need use cases for the workflow around tax firms, I usually see a lot of workflow capabilities and operations, and, you know, bringing in emails, summarizing things for clients, you know, because you can’t give it the sensitive information. Building out the depreciation tables, things like that, I’ve seen, but, like, Canopy, TaxDome, they’re really coming along on their tax products.
Seth Fineberg: Yeah, there’s I mean, AI is getting built into nearly everything that you use, but outside of that, you know, particularly with the popularity of Claude and GPT, you know, obviously some folks going, well, is there any in particular better suited for accounting? I mean, I’m not hearing anything one side or the other myself. I mean, both of them allow you to create in different ways, yes?
Jan Haugo: 100% agree, yeah, yep. Yeah.
Seth Fineberg: And apart from them, any other AI tools that they could try hands-on? Again, I know that you mentioned in your presentation – you mentioned Gemini, particularly within the Google environment and Copilot within Microsoft.
Jan Haugo: Yeah, I would say a great way to start is to start using the tools within using the AI within the tools. So, like, I use Notion. Notion has an AI inside of it; start using that, because it’s a great, safe environment. I think Asana’s got some, so pretty much all of the tools are starting to build AI in, and that’s a great way for you to get comfortable with it.
And then when you feel like you want to venture out and go out into your own workspace, use Gemini or Microsoft and try to build some things in there. I can build the bookkeeper buddy in any of those environments. I can build my sales agent in any of those environments.
So, get comfortable in the environment you’re living in, and start to build out a few tools. The biggest and easiest way is two things. You ask yourself, what is the most painful thing I have to do, and what’s the most repetitive? And focus on those two, and that will help guide you into building.
Seth Fineberg: They could be one and the same. That’s where we’re at right now. I had heard the analogy that I often hear is that AI right now, in its current iterations, is a great band-aid for a lot of things, and it’s probably the best band-aid we’ve ever had for so many things.
But even over and above that, the things that you can create are so personal, rather than waiting for some vendor to come out with something that you hadn’t thought of yourself.
This environment allows you to kind of build I mean, look, there were just two things in the past hour that Jan has shared with everyone here that she built for herself, her firm, her people, the work that she does, because she felt that she needed it.
Jan Haugo: Yeah, but look at our industry as a whole, and I know we have to go, but what are accountants good at? Solving problems, right? So what is the best tool that they have to solve a problem? AI.
Seth Fineberg: Exactly, and it really is some of the best that we have out there. I’m sorry if we couldn’t get to every single one of your questions out there, but I do thank everyone for their time. Jan, anyway, you know, do you want people to reach out to you with any further questions, or are you open to that? I don’t want to put you on the spot.
Jan Haugo: You’re no worries. You can always reach me on LinkedIn, and then you can always email me at [email protected]. And, you know, honestly, the people at Ace Cloud Hosting have been a fabulous host, and they can also get a hold of me and, you know, take advantage of the offers that we have, because, to be quite honest with you, there’s so much out there, but I’m more than happy to answer questions and, noodle on things with you.
Seth Fineberg: Excellent. Well, I look forward to talking to you at another time in the hopefully not-so-distant future. Thank you all for taking time out of your days to sit here with Jan and me, and of course, thank you as always, to our host, Ace Cloud Hosting.
Definitely reach out to them for any of your managed cloud services or managed security needs. If you have not already, I think you’ll be very impressed by what they have, as well, to keep all of the tools that you use in a very secure and usable environment. Take care and have a great rest of your day.
Jan Haugo: Thank you so much.
Seth Fineberg: Thank you so much, everyone.
Subscribe to Ace Cloud Hosting today to receive expert content, practical advice, and thought-leadership insights that help businesses stay competitive, secure, and future-ready.
