The Modern CPA Success Show: Episode 139

Still stuck managing your finances in spreadsheets? Today’s best CFOs are using AI to streamline operations, fix data issues faster, and boost cash flow visibility—without increasing overhead.

In this episode of The Modern CPA Success Show, we talk with Ahikam Kaufman, co-founder and CEO of SafeBooks AI. With decades of Fintech and SaaS experience, Ahikam breaks down how AI is transforming financial data governance, automating repetitive accounting tasks, and making real-time financial insights a reality.

Tom Wadelton: [00:00:07] I’m happy to be joining you again today for another episode of the Modern CPA Success Show. I’m Tom Waddleton. I’m one of the virtual CFOs at Anders Virtual CFO. Co-hosting today along with Adam Hale and Adam leads our virtual Cfo practice. Adam, welcome. We have a guest today, Ahikam Kaufman. I apologize for mispronouncing that. Ahikom is the co-founder and CEO of SafeBooks AI. So they’re revolutionizing financial data governance with AI-driven solutions. He’s got over 20 years worth of experience, including leadership at major firms like Intuit and also guiding startup firms. So we’re really looking forward to a conversation about how tech is transforming finance on. I come welcome Hey, Tom and Adam, thank you for having me this morning, looking forward for our conversation. As we are, so I’m really interested, what led you to the decision to start SafeBooks AI after doing so many different things during your career?  [00:01:06][59.1]

Ahikam Kaufman: [00:01:08] Grew up in the office of the CFO and what I realized is while there’s a lot of focus on the operational layer, on the reporting layer, I think we all rely on data, you know, daily work, whether it’s even a small business or a large company. The challenge is when you start to get to become like a mid-size business or large company, then you have a lot of disparate systems in the office of the CFO and the ability to monitor the data. Across the systems, especially where every single transaction would typically be processed by multiple systems becomes a challenge. Why do you need to monitor the transaction? You wanna be able to monitor them for data integrity to make sure the data is okay. You wanna monitor them for control purposes, and you wanna be to monitor for documentation purposes because that’s part of your closed process. And it’s really, really tough. I’d like to think there’s a gap between the need of us as accountants to sign off on the data and our ability to do that with a complex data environment. And that’s what we aim to solve.  [00:02:17][68.9]

Tom Wadelton: [00:02:19] That’s great. And I hear a lot of mention of AI. As you know, it’s hard to go to any conference or read anything that talks about AI things. What I like, it is different. I’m looking forward to this conversation as most of it is efficiency and how AI can reduce the people involved in work, all sorts of things like that. I hear very little about financial data governance. So I’m curious that focus area that you have because that’s not what usually people are talking about with AI.  [00:02:44][25.3]

Ahikam Kaufman: [00:02:46] Totally. Actually, if you Google financial data governance, we’re likely going to be the first unsponsored result. I’m sorry, although Gartner declared that financial data and analytics governance is the top priority for CFOs in 2025. It’s the number one priority. And the reason is, when you run a organization in a company because you know you have complex processes, a lot of people, you need to develop like a control muscle. So it’s not just about transacting or doing your job. It’s also being able to reconcile the data to find mistakes. So you need have control and governance across that. And I’d like to think that until now, I think you guys had a session in your podcast a few weeks ago about Excel, 30 years after. I think a lot of the processes in the office of the CFO have changed dramatically and people still rely on Excel to augment the gap between their ability to consolidate data or look at data from different systems, different processes and just the lack of technology to do so. So the easiest platform, if you put everything on Excel, that’s easily manipulated. You can automate all kinds of things, but it’s not repetitive. It’s not scalable. And I’d like to think with AI, we can achieve a couple of things. There are new technologies today that allows us to do things we haven’t been able to do in the past, in terms of our ability to monitor and govern data. And AI will allow us to smartly automate activities that in the passed require just a human intervention. So I’d like to see If we would look today on many of the tasks in the office of the CFO, you can automate them with agentic AI, meaning smart AI that looks at the issue and decide the course of action. We have to augment it, right? We can’t let it hallucinate because it’s still the office CFO. This is why I’d like to think that’s the risk in AI that we’re trying to mitigate. But I think AI allows us to. To look at data in a different way, complex data problems, and also be able to automate repetitive activities that people in the office of the CFO for many, many years had to do until today manually or using spreadsheets and all of that. So a lot of the work hasn’t really materially changed. And I’d like to think now we have a good chance to automate it, make people feel better about the data. Make them focus on more strategic tasks, as opposed to repetitive, specific, maybe more complex boring tasks and do more with less.  [00:05:50][184.4]

Adam Hale: [00:05:52] Let me ask you this, I was actually just having this debate this morning, because we’re looking at deploying some AI tools, what we’re hoping to be AI tools. Right now, it still feels very much like a buzzword, feels like a little bit of advanced automation, which, as you know, automation is not, you know AI, but in fact, I heard to quote one AI person, Automation is AI’s dumb cousin. And I was like, that makes more sense, I actually like that. But we were kind of talking today about, so in our world, in the CAS practice, whenever you’re doing client advisory services, you’re not working with, we don’t work at GM, so we’re not like full-time CFOs, and we’re working with 9 billion records. What we’re doing is we’re typically working with small to medium sized businesses that to have a hundred to… 500 transactions a month generally, they can have 1,000 or 2,000, but we’re trying to jump from client to client to clients. Those clients are going to have different systems, they’re going to different inputs. I’m not necessarily needing to, you know, they’re not always going to have the same general ledger or accounting software, but they’re also going to have different time plate, you know, time entries and different software. There’s going to be like this combination of in order to get to that real time reporting that you’re talking about and be able to do that more valuable stuff, I’ve got to be able to connect three or four different data sources and pull it all together into one data lake. Um, whenever it comes to automation, one of the biggest challenges is that, uh, in order for automation to even remotely work, everything has to be completely standardized. My only problem with that is now I work with 35 clients. And they all have a different combination of GLs and time tracking and CRM and all these other kind of things so what where we become super and efficient whenever we’re picking up clients and we’re working from a fractional cfo perspective is like if i have to spend. Days and weeks and months and have a programmer standardize all their data in order to get the reports that becomes less and less efficient. My understanding was like true AI agent learning is that we can start getting away from the need for all the standardization and we can basically just like kind of dump the information into this gigantic data lake. And that the AI agent has the ability to standardize or pull the information and learn and do those kinds of things. Is that true or are we still in a place where we have to do heavy standardization in order to get this reporting to work out?  [00:08:40][168.8]

Ahikam Kaufman: [00:08:42] Oh, Tom, I think your question is very complex. It’s a really good question. And I’ll try to break it into a few pieces and divide and conquer, because otherwise, it’s a complex question. So first, you made a comment about whether AI is a buzzword or not. And I fully subscribe to that. What I would tell you is the following, right? And we have to kind of control the buzz around this. And just to give you an example, if that’s okay, if I’m a programmer and I write a piece of code and I’d like to test that code, right? Because most likely OpenAI have seen at least half of the entire code in the entire world, then the OpenAI ability to respond to my query about the quality of the code or whether I have bugs is pretty good, right. And that’s like the stage AI is now. It can help you with tasks. Maybe it’s like a dog, you throw a stick at a dog and he brings it back to you. So very structural tasks, whether they’re on structural data or unstructured data. It’s not scalable and automatable, right? So you can give him task. It’s more like a knowledge base, a human-based knowledge base. It’s maybe like how we viewed Google 25 years ago when it was first exposed, right. All of a sudden with your fingertips, you can access all the data in the world, right, so now you can ask a system question and ask it to compare. A set of data you provided versus all the data in the world they’ve seen. So that’s a very structured process. I think that’s where we’re at right now. Our job is to take these capabilities and tailor them and augment them into a structured process that can actually work. You mentioned, so for example, if we can build capabilities around document processing, or document processing or understanding or whatever, so if I can take AI to view an unstructured piece of data, like a document, pull out data which is relevant to me as a product maker, whether it’s like invoice data. Or billing data or whatever, and I can control the outcome. I can verify that it correctly pulled the date, the amount, right? So that can really work really, really well. However, again, it requires training and customization. I’d like to think you brought up another point where even if you have like an AI powered solution, it’s not a one size fits all. And I think that’s like part of the problem we’re tackling now. Because we are dealing with larger companies, we do have to take whatever we built, the technology, and tailor it to the customer, to its systems, to its processes, and then it can work. But to take an off-the-shelf product, AI-powered, and be able to work with dozens or hundreds of customers without any customization or training, I think you’re right. It’s not there. Maybe you can do certain things like document processing, things around that. But for most of what AI can provide now, it’s not going to be a good solution. Again, a year from now, it’s going to totally different. Plus the fact that AI hallucinates, which we have to control that. So AI many times would fake an answer. So AI today is not programmed to say, I don’t know. It will always give you an answer which you need to check. This is why, this is why in our solution, and I just don’t wanna deviate too much from your examples, but in our solutions, we contain the AI to do very certain things that it can do really well and things that I can also. And control the quality of the outcome, as opposed to just take a bunch of data, throw it in the AI, and ask it to give me something. That’s not gonna work, and I think you’re very much right there. However, you do have capabilities, like, again, to be able to decipher unstructured data, to be to read the document, where I can augment existing technology with the capabilities from AI and still control the outcome. And that’s what we’re doing. So a long answer to a long question, but I think you’re right. Certain things can work really well. Certain things, it’s still a buzzword.  [00:13:40][298.3]

Adam Hale: [00:13:43] Yeah, I mean, that’s kind of where we are. I think there’s a lot of AI out there right now that’s reading transaction data. They’re helping with bank reconciliations, credit card reconciliarations. They seem to have that down pretty good, that task, and that seems to be pretty universal. They can bring the data down and do those but i think where you know in our profession anyway where you could really save a lot of time and effort and energy is if you could automate the month and close. So if you can pull together all the worksheets all the work papers do a lot other data validation and set that stuff up i think that’s where we have the biggest opportunity to win. Um, and trying to work with numerous AI providers to, you know, kind of come up with that solution. But again, the problem for us is, is that the data is unstructured. Um, And you could structure the data, but again, if you’re doing it for every single client and you’re.  [00:14:42][58.4]

Ahikam Kaufman: [00:14:44] It’s just, it’s not cost-effective yet to do it. It’s most effective to do for a large company. It’s not called right. And it doesn’t walk out of the box. So because it’s a self-serve and that’s exactly, I think where you are right now. I do things that, you know, for example, maybe, you know, tax questions, tax advice, you know, sophisticated search, though. Multiple things that they, I can do very well today. Right on multi on the sophisticated search, right? Because when you go to Google, it gives you a bunch of links, right? Right now, we have Gemini, but in AI, you can ask the question and you can asked it again, it’s a very unstructured process. But it can be like, I don’t know, maybe a tax expert or whatever. Again, you can process documents, you could do very limited things within the world of accounting today. But in order to have a more sophisticated solution, I’d like to think that what we’re doing, which is very much tailored into every company process, that’s the way to go now. And I agree, it’s not cost effective yet for small businesses.  [00:15:54][70.6]

Tom Wadelton: [00:15:58] So, Ahiqam, at the engagements that you work on, you had said it’s mid-size companies. Do you want to help us understand kind of how that works? It sounds like you’re going in and doing some customization and say, okay, for this company and this set of tools, but we’d love to hear what that looks like when you engage with someone.  [00:16:12][14.4]

Ahikam Kaufman: [00:16:12] Right. Right. And I think a lot of it kind of Tom touched on. So I think so basically what we’re doing is when you run a large business, you need to have controls. You need to check things on a daily, weekly, monthly, quarterly, and sometimes annually basis or multiple times across the year. And those are like specific repetitive tasks, which requires you to pull data from different sources, analyze the data, check it, remediate it, and document it. And that’s what SafeBooks does. So for example, if you have like a complex order to cache process, which includes like, you know, documentation. Orders, sales order documentation, which is processed by a CRR, which then goes to a billing system, which goes to a payment pathway, which then goes to an ERP. And then you have like refunds and RMA and returns and things like that. And you need to reconcile all that data and you need to check all the data. You need to check that certain things happen. We can automate all of that in a structured environment, which is like a given customer. Right? The data structure, but it’s a structured environment. We can do it like there’s a one size fits all. I’d like to think, uh, uh maybe 10, 15 years, 15 years ago, a company started the cold Palantir, which is a very successful penny. And I’d love to think we’re very much doing so they created this platform and, but you have to build use cases for every customer, which are tailor made for them, for them. So our ability, to aggregate all the data from the disparate system, from the different sources, mesh it together using our technology, create like what we call is a single graph database, which interlinks all the day across the system, as if it was a single system, allows us not only to check, validate the data, but also apply AI on top of the data now that all the date is organized in a way that can replace manual tasks that people are doing today, mostly using spreadsheets, including fully automating wallpapers. So every wallpaper, what is a wallpaper? Wallpaper is when someone takes a bunch of data, whatever, document, captures it, and then apply a certain logic and he checks all kinds of things and want to make sure. And then when everything is fully checked, he documents it in that piece of paper. We are fully automating that again for every given customer. And I’d like to think that by doing that, we allow people to save a lot of time associated with pulling data from different sources, but not just that. We allow people too. First of all, remediate issues in real time, because in order to remediated, you have to know what the problem is to know where the problem is, you to do all these processes, which takes you a lot of time. All of a sudden it’s now prepared by the machine. The other thing is we expedite the close and audit process because all the work paper used to spend a couple of days on each to prepare, they’re now prepared automatically. The other things is financial data is money. If you have a mistake, it’s not just data mistake. It could also be money you leave on the table. Last week, last week we told a customer that due to data issues, they over accounted for their revenues by like six million dollars per month, meaning they had certain things which were not fully reconciled, which caused them to think that their business, unfortunately, was bigger than what it really was. But when you have investors, when you have like. When you need to report it to tax, it’s very bad to make these mistakes. If you need pay for example, tax for this, for the data. In another, with another customer, we actually were able to find that they under build their customers, right? So think about businesses who will be like, by the way, sometimes you have small businesses like that. They need to build 20,000 customers a month. You have mistakes. Our ability to automate that data governance, that process these processes. And make sure that every single week you can make sure that all your customers will be on time, allows you to kind of get that value. And that’s like some of the things you’ll do.  [00:21:05][292.3]

Tom Wadelton: [00:21:06] I’m curious in the example where you said that they weren’t billing the customers. You said we found what did that look like that the tool itself was coming and saying, hey, here’s something I noticed or you guys notice and pass that along to the customer. What? How did that play out?  [00:21:21][15.0]

Ahikam Kaufman: [00:21:23] That’s a great question. So think about a company that transacts through multiple platforms. Again, it’s not the case, but I’ll give you an example, maybe in your clientele. Think about that you have a customer or a client that operates through, let’s say, Shopify or other marketplaces, right? And then, for whatever reason, if everything is manual and you will not fully sink. With all the marketplaces, you can have a situation where you process the transaction, you even ship the product, but you didn’t build for that product because again, I’m talking about large numbers. So imagine you have like 5,000 transactions a month with Shopify. So in our case, we had, we have a customer who transact through a marketplace. Because everything is manual, people were not really agile enough to follow up on all the transactions that were processed through the marketplace. It’s a software business, so you don’t actually ship anything physically, but you have to build and the customer get the value. The customer got the value, but they didn’t build. So we actually found that they underbid their customers for like $2 million. For the last few seconds as a result. The way the experience looks like is imagine when we pull the data from the marketplaces, from your ERP, from your billing system, from your payment gateway, we create processes or we create like agentic AI powered controls that checks all of them. And the way you consume it is through the applications that we provide or through work paper. So if you produce that wallpaper that is responsible for reconciling all your business then we would flag in red these conductance where we can’t identify the bill, the invoice, or sometimes we can identify a credit bill, it could also be the opposite case, it’s also relevant. We couldn’t identify the building. And then they investigate that, because it’s their own data. It’s easy for them to investigate. And they found out that they don’t have, for whatever reason, that the invoice was not processed in the ELP.  [00:23:50][146.8]

Tom Wadelton: [00:23:52] Those are great examples. So if you think about smaller companies and many of our listeners are accountants who do advisory service but they tend to be on the smaller companies, what advice would you give because we’re all talking about AI, we’re doing different things but how could what you’ve learned with SafeBooks apply to these things or maybe SafeBooks applies too. But what would you have us do? Because we also have. Accounting systems and other systems like think of like a bill.com plus cookbooks online. Maybe other tools like a time entry kind of system  [00:24:24][32.2]

Ahikam Kaufman: [00:24:25] Yeah, definitely. I think I’ll say the following. First of all, whatever could be automated should be automated. So automation and I always a friend, a good friend send me a few months ago, this caricature of like people trying to move something really heavy like thousands of years ago using a car with a squared wheel and someone was standing on the side and selling a rounded wheel and they say, we don’t have time for this. So most likely, most of us would say, we don’ have time to invest in automation. But I think to carve out time to invest in the automation of like in processing the invoices, even if, even if your job is only cut by a half, which means like you find a platform which extracts all the relevant data, you need to approve it. But instead of just doing it yourself, you just confirm it and you’ll And it cuts half of the time. So I think investing in automation, like you guys mentioned, on reconciliation, on document processing is essential to make your margins better, to make you less error-prone, and all of that. I’d like to think the other thing, even if you’re a small business, invest in tools, right? The way I, the analogy I use with people that are not always coming from the industry. All of us, I think, apply discretion where we swipe our credit card. I think a fraction of us actually step down at the end of the month, read the statement, and either like they used to do in the old times, reconcile the checkbook or just please that the statement doesn’t have any, that’s control. Every business should have control. And our ability to be able to apply even the minimum amount of control, even for small business. Whether it’s around checking the bank reconciliation, checking inventory items against what was sold but was not sold, reconciling inventory, things like that. I think it’s important. I don’t think it will be automated any time soon for a small business. But being able to develop that control arm, not just the transactional arm, and be able to say, OK, where my risk is in my operation? And again, I had small business. Where is my risk and what kind of process I can apply even like on a small spreadsheet that could make sure that I didn’t miss anything, I think it’s essential.  [00:27:05][159.6]

Adam Hale: [00:27:08] Yeah, no, that makes a lot of sense. And then so, you know, is there anything out there like fintech or AI trend that you’re watching really closely that you think that firm owners should pay more attention to?  [00:27:19][11.5]

Ahikam Kaufman: [00:27:25] Um, um, you know, uh, hopefully, you know, the whole trend around open banking, which we dealt with a lot at Intuit and Intuit obviously caters to your world better than I am. Um, uh is really, really important to be able to get all the bank data. We’re now dealing with a very large bank and it’s really complex to get the data for one of our customers. So I think the the being able to share data from third parties, whether it’s like your build.com platform, it can help a lot when you’re trying to automate if one of your businesses grows a little bit, being able pull data from build. Com, being able put data from your bank and maybe create processes that will reconcile things automatically, I think it’s important. That’s really what I can, that’s like a trend we’re tracking right now. When you select a system, make sure that they have like an open, they have a positive approach for open data, right? So if you select the system for your customer, whether it’s a build.com or on the van or whatever, just make sure that their data is extractable, is accessible. So in the future, six, 12 months from now, you have another system that can plug into these systems to help you automate another action that vendor you selected offers that access to data. And because today, a lot of the time, some of the challenges are every system, even like the big ERPs, they’re solely focused on being able to push the envelope, move the transaction from point A to point B. You know, they, they less invest in like, how do you can make sure that that transaction across other systems resonates and, and, um, the inability today of companies to apply controls across systems, across system, not in a single system is because each system it’s like, um a walled garden to a, to some extent, so being able to choose vendors that allows you access to their data. So you can in the future interact with that data automatically is important.  [00:29:45][139.8]

Adam Hale: [00:29:46] Yeah, no, I agree. I think one of the challenges that we see with that is at least like even in our service agreements is clients more and more asking, okay, what are you connecting to? And how are you using the data? And then, you know, from us just making sure that we’re also securing that data is kind of a huge thing there. Any tips or tricks there for things that we should be aware of and consider from a security standpoint?  [00:30:14][27.4]

Ahikam Kaufman: [00:30:15] Uh, I would, I will say the following first, all of our data is in the cloud. I’m pretty sure that even most of your clients, when you pull the data, their data is already in the Cloud, but it’s just stored by whether it’s QuickBooks online or Sage or whatever, all these solutions. Um, I think, I see you need to have like, um, You know, because a lot of the work you do is with staff members, I’d like to assume, I think you really need to have like, to develop like a security program, which includes the following. And I’m not a security expert, right? Try to minimize, uh, the exchange of data over like non-secure platforms, email included, if possible. Try to make sure you have a storage place for every firm. Like you can use Google cloud or whatever, but try to use like. Don’t don’t store customer clients data on like hard drives or things like that, try to put it on our secured cloud environment, because those are secured and they’re like, you know, they replicate the data and all of that, and try to also have like, a training for your staff members. And the importance of training, which we also do internally in the company is to just see the mindset of security. You know, don’t click if you, God forbid, you have customer data on your laptop and you click on a suspicious link, you can expose all of that data imminently, right? Having the mindset of what to do and not to do for a secured environment with your staff members is also important. So we can never overestimate the importance of refreshing those rules of engagement and what to and not do in order to secure your client’s data. Try to save to the extent possible as much private and less private data as possible like social security numbers, payroll data, that requires a special scrutiny also, maybe limit the amount of people that deal with it and yeah, it’s gone.  [00:32:34][138.6]

Tom Wadelton: [00:32:36] I’m going to go back for just a second when you talked about having the systems you choose make sure they have open data. If many of the CPAs are not terribly technical, what kind of things would they look for to see if it does have open? Are these things like we have API connections and things like that or is it something  [00:32:53][16.7]

Ahikam Kaufman: [00:32:53] Yeah, API connections. Unfortunately, and I don’t want to name names, but unfortunately, we’re dealing now with two large companies who selected a vendor, a payment vendor. And in two separate cases, the payment vendor, which is again, two different companies, two different payment vendors, they don’t also access to the data. So now the large company would like to automate some controls. You know, the office is closed. We can’t access the data. So I’d like to think that even if you do a simple today, by the way, you don’t need to do a Google search. You ask open AI, right? That’s the, here is an example of using AI. You go to open AI and say, I’m an accountant who supports a small business owner. Who would like to purchase the system, I want to make sure we would have access to the data, so an API if needed, and you name the name, and whether you use Perplexity or OpenAI, it will give you the answer. So you don’t need to be an expert. You can trust the OpenAI or Perplexities for that.  [00:34:10][76.6]

Tom Wadelton: [00:34:11] What’s your thought about the AI tools that are happening within software? Like Intuit has its own AI and build.com has AI, things like that. Most of the examples that you’ve been given have been saying, okay, let’s look across systems. I’d love to hear a little bit more about whether you think they’re good. And then I guess the limitations of AI within.  [00:34:30][19.2]

Ahikam Kaufman: [00:34:32] I actually think, as I mentioned before, that the use of structured AI within a system should work really, really well. By the way, people are not aware. We can talk more about it. I’ll give you another example, which is not relevant for accounting. I’ll just give you an example, but I’d like to think, for example, one really painful issue is to avoid double payments. Okay, create a lot of controls for large companies to avoid double payments because You have segregation of duties, and you have a lot of chefs in the kitchen in a large company. But I think for a small business, even when I was involved in small businesses, people can make mistakes. So the ability to use AI to identify mis-reconciliation or double payments or things like that, I think in products like Intuit or Bill is essential, is essential. Meaning to help you avoid mistakes. In that respect, and I think people are not really mindful about that, Google recently added the Gemini AI to all of its products. So right now, imagine, and I want to just seed the idea in people’s mind. Let’s say you look for, you have a thread with a client about a topic. So, and you want to. You want to check where did you land with that thread, right? Because maybe it ended halfway through the finish line. You didn’t get to the finish, right. What we used to do, we used to go to Google and search in our email and see all the thread and start reading it. Today, you can ask Gemini the question. You can ask Geminis within your Gmail, Hey, I had this thread with the client ABC. Did we agree on an action ID mode? It’s that crazy. You know, the ability to, meaning what I think your listeners should adopt is, if I used to do something specifically and I used the work hard, if there is an AI component needed in whatever platform I’m using, I’m trying to use that first when you have a question, maybe the AI can give you the answer. Again, it’s not a software process, but it can save you time and effort.  [00:36:52][140.7]

Tom Wadelton: [00:36:55] OK, that’s a good example. So, if you’re a CFO and I heard you say that…  [00:37:01][6.4]

Ahikam Kaufman: [00:37:01] By the way, you can, for example, let’s say you have a meeting with a client, you could go today to Google, I guess, I think, inside Google Calendar with Gemini and tell him, hey, I want to schedule a follow-up meeting with my client on a monthly basis, this and that client go over certain things. And Gemini will do it for you. And it will schedule the meeting and it will check that you got acceptance and all of that. It’s like daily tasks that are now embedded into the platform we use every day, which you can really, really leverage and save maybe half of the time you would spend on that task before.  [00:37:39][37.8]

Tom Wadelton: [00:37:42] Okay that’s a good example. So if we bring this back to what you think our listeners should do and you had mentioned that Gardner said financial data and governance is the top CFO priority for this year so if someone’s giving advisory services and maybe they haven’t really thought about governance and things before what would you suggest hey here’s a really important thing that that you should focus on.  [00:38:05][22.9]

Ahikam Kaufman: [00:38:06] I actually, we’re working with a few advisory firms on being able to offer automation for staff which previously they consulted to clients and ended up with like a protocol of activities they need to do manually on a monthly basis, right? So a lot of companies, they use advisory services to say, okay, well, my risks are and then what controls should I put in place? And then these controls are turned into like monthly manual activities. And today you can provide it with automation. So I urge every, if you have any listeners that is working with like a midsize or a large company and they can consider, we will be happy to show them or talk to them about how these devices, which previously are turned into like manual tasks and spreadsheets that you need to produce. Could now be automated, letting the customer, first, not spend the time on that task, second, get the answer imminently in real time so they can remediate as opposed to. So I’d like to think there is also a portion. And I think that connects all the dots around the use of AI. We need to think the way we need to think when we wake up in the morning is that everything we’re doing today could at least be possibly be done better and easier. We just need to force ourselves not to tell the guys who sell us the rounded wheel we don’t have time, but to spend a few minutes to see how we can use Gemini, leverage Gemini in the calendar, leverage Gemini on the Gmail, or if you run a larger business leverage a tool like or a solution like savebooks in your daily activity. It requires a little bit of investment but it can dramatically change how we operate on a daily basis.  [00:40:00][114.2]

Adam Hale: [00:40:02] I think you said something earlier on that that really resonated and I’ve heard before is um don’t get the you know basically don’t get the shiny object syndrome there’s a lot of different tools out there but really just start with understanding what the problem is first and then attack it from that direction so what is the one thing that would you know for you you label it risk what’s the one things that it risk the most you know and for me whenever I was looking at it it’s like what’s going to time. If this was automated, if AI could assist. And then we went to market trying to find that solution. So for us, it’s that month-end close is what we’re ultimately looking for. But I think just that perspective is good, because to your point, there’s a million different tools out there. You’ve got to first figure out what you’re trying to solve before you go looking for the right tool.  [00:40:53][50.9]

Tom Wadelton: [00:40:56] Yeah, that makes a lot of sense. Adam, that sounds like your big takeaway. Am I right on that? If you think that. Yeah, yeah, absolutely, yeah. Yeah, I think for mine and I said it early on would be so much of what we’re hearing from AI is what you’re doing today, you can just speed up and let me show you how much easier it is. I don’t think I’ve heard much about from a controls perspective and that was really valuable.  [00:41:18][22.5]

Ahikam Kaufman: [00:41:20] Right. I think the number one KPI would be that your listeners and colleagues would ask themselves, I want to make sure that me and my team are using AI on a daily basis. So if they have a tax question, they have like back to your point about a set of data they want to look at, or again, doing a search in emails to find a thread. If you used AI on daily basis. And even just scratching the surface, it means you’re making progress. And I think that’s what we should kind of educate ourselves to do more and more and, and see if we can get used to that. And then it would lead to good things down the road.  [00:42:03][43.0]

Tom Wadelton: [00:42:03] I think that’s a great tip just getting comfortable with all the different possible things that could be done. You can learn so much  [00:42:09][6.0]

Ahikam Kaufman: [00:42:10] You know, I’ll tell you, it’s interesting. I don’t want to name names, but in our sales process, I asked my team, do you use an eye and they said no. And what I told them, listen, today we have prospects, okay, and before the call with the prospect, we want to know everything we want to know we can know about a prospect, about a company. Instead of searching in Google, Searching in, let’s say, open-air and perplexity. Try to force yourself to use AI. It’s easier because once you do that and once they did that, they said, wow, that’s great because they just give him the answer. It’s like an admin that did a small investigation online and gave you all the risk to know about that company. Especially if it’s a public company, but even private company. So just force yourself, train yourself to, to use it instead of like the normal channels. And like Google and others. And you’ll see it will give you, it will save you time and it will seed this sense of automations that can lead to good things down the road. I think that’s an excellent tip.  [00:43:28][78.5]

Tom Wadelton: [00:43:29] It’s probably a great place for us to close. Yeah, great. Hi, Kam. This has really been helpful. Thank you. I kind of opened up some different ways of thinking about things and love the SafeBooks product and the thoughts you put into that. But also, even for people who work with smaller companies, just the concepts that can flow through. So really appreciate you coming on and teaching our audience.  [00:43:48][19.2]

Ahikam Kaufman: [00:43:50] The way your listeners should think is that OpenAI was trained on all the data in the Part of the data is small businesses. They see a metric in a small business, they can compare it to other businesses. They just need to ask OpenAI because it’s like this consultant to work with all the companies in the world. Every question you try to ask yourself, instead of working out to answer it, try to check online. Maybe it can give you some of that answer or a way to… A way to approach the question. So, yeah. Thank you very much. Thank you. I’m not discriminating, as you said. OK, thank you, Tom. Thank you, Alex. Yes. It’s been great.  [00:43:50][0.0]