#556 The CFO’s New Mandate: Ahikam Kaufman on AI, Financial Governance, and Real-Time Truth

[00:00:00] 

Mehmet: Hello and welcome back to a representative of the CTO Show with Mead today. I'm very pleased. Joining me, Ahikam Kaufman, he's co-founder and CEO of SafeBooks ai. Ahikam the way I love to do it. I keep it to my guests to introduce themselves. I [00:01:00] know you have like a very rich, very, um, uh, you know.

Inspiring, I would say, uh, career. So, but as I said, I like my guests to introduce themselves. So tell us more about you, your background, your journey, and then we gonna start to talk, uh, about many important topics today, mainly in finance, in in, uh, compliance, and of course little bit of ai. So the floor is yours.

Ahikam: It's, uh, really an honor to be here. I know you have a very popular podcast. Um, thank you. I know you have listeners from all over the world. Uh, for me personally, I started my career as a finance person in a Bay Area startup. Uh, so I was actually, um, I, I, um, I learned accounting. I worked for, for Anderson for four years, and then I went to a startup because, um, uh, about 25 years ago, tech started to become more popular, um, eh, and, uh, [00:02:00] uh, I found myself, um, you know, serving as a young CFO for a software company.

Which, um, uh, lucky enough, after a couple of years was acquired by a larger public company. And then, um, uh, I moved to that public company and decided to stay to learn how big company operates, especially in finance. And, uh, the company has been very successful and that company has been acquired by Hewlett Packard, by hp.

And, um, after, uh, after being acquired twice, I decided that, uh, because I was in Silicon Valley. You know, uh, maybe it, uh, it would be, would've been appropriate not just to pursue a finance career, but maybe a more entrepreneurial career. And then I started my entrepreneurial career together with a friend.

We started a, uh, a B2C payment company. Uh, at that time, uh, the world has switched from Blackberry to iPhone. So we decided we, we started [00:03:00] to, uh, distribute. Our product, our application on the iPhone. We actually developed the iPhone application before we even had an iPhone because during that time, the early years of iPhone, you couldn't even buy iPhone easily.

There wasn't enough. So, um, uh, that was funny. But, uh, we created a, a payment application, which became very successful. And we got acquired by a, a large company called Intuit. And at Intuit I learned a lot of things. And, um, that's eventually how I got into Safebooks. So, uh, served like about 10 years as a finance guy.

And then, uh, uh, over the last 15 years in Entrepreneur, and two years ago, two and a half years ago, I decided to do something that combines both my career as an entrepreneur and as a finance person and do something in cope with finance, which I think is starving for innovation. And I'm happy to see that today.

There's a lot [00:04:00] of innovation, the office of the CFO, but the task hands are complex and that's what I hope we can cover, uh, um, to some extent today. So thank you for having me. 

Mehmet: Oh, it's my pleasure. And you know, thank you for making today come. So, you know, maybe a traditional question for you, I gonna. You know, start with, with the, you know, the motives to start Safebooks, right?

So, and you work yourself, uh, you know, you start as an accountant and then as a CFO for large, for both startups and large companies. So I. I'm interested always in the original stories, and of course, like you have, you, you come from, as we call it, like a practitioner background. So, so you, you've seen a lot and, and you know, a lot of the space.

So what was, you know, you know, the main thing that you felt that it's, say kind of a. Big challenge. Some people they like to call it [00:05:00] blind spot. You know, especially when it comes to, to financial, uh, data governance in, in, in, in companies that pushed you to say, okay, there is an opportunity there, there is a huge problem untouched, or maybe there are like some solutions that we can do better.

I'm always, you know, keen to, to, to know these original stories, Ika. 

Ahikam: Okay. So, uh, I think it's a great question and let me try and explain that and I hope. For those in the audience who doesn't put the finance background will be, I'll be, I'll be, uh, uh, descriptive enough and, uh, detailed enough to explain.

Um, every company has a finance organization, whether a small or big or fully. Everyone knows that in large companies you have two fundamental problems when it comes to. Uh, corporate finance, two, two fundamental problems and keep me honest still, I don't explain it. Sure. The first problem is that unlike [00:06:00] many parts of the organization, maybe similar to marketing, but the, the people that needs to do most of the work are accountants.

But most of the work is around data. When you are working in engineering, when you are working in, uh, DevOps, when you walk in security, you deal with data and with systems, but you are like an engineer. You are technical guy. When you deal with finance, you are an accountant. You know, the accounting rules, you know, um, the um, uh, uh, how to put together numbers and business stories.

You're not a data expert for the most part, but what you need to deal with is data. So there's a gap between what you were trained to do and what you need to do in reality, right? It's almost like, you know, someone wants to start a business even as a lawyer, okay? So it gets very excited about that, but the first day he needs to deal with.

The salaries office [00:07:00] people, and, you know, the office administration and all kind of things which are not related to his core expertise, but as like, as a, as a small business owner, you need to deal with, same in finance, you need to deal with data, but you're an accountant. That's one gap. Okay. Does that make sense?

Mehmet: Yes. 

Ahikam: The other gap is that, uh, like many other parts of the organization. Your data travel between the disparate systems. Okay. So, you know, and that is a problem or that is like, um, uh, uh, uh, something that happens in many parts of the organization, in engineering, in marketing, in sales. No one operates in our single system.

You, you know, think about, uh, for example, soy tool, because there are many similarities to what we do in security. Think about the office of the ciso, right? 

Mehmet: Mm-hmm. So 

Ahikam: he has probably 40, [00:08:00] 50, 60 applications that they're running, and they need to navigate between all of them to manage the priorities and what to tackle first.

And, and all of that same in engineering, same in marketing. It's really that one organization have one system. Okay. In finance, because you have multiple systems, the problem amplifies because, because when you are a public company or when you are a regulated company, or when you owned by large shareholders, the compliance requires you to have like an audit trail and governance for the data.

Right. So it means that if you have like a sale transaction that started as a quote, as a commercial quote, and then it went to your CRM, and then it went to your billing, and then it went to your ELP, and then it went to your bank, and then maybe you collected it, you have to be able to track it across all the [00:09:00] systems.

So. In marketing, it's also important to bring all the data from all the systems, but not necessarily the same transaction goes through all the systems and you don't have compliance in marketing. In finance, you have compliance. And compliance requires you that all of the data will be governed and consistent because you don't wanna be in a situation where you report collection on deals that are not related to the quarter, or deals that were like, meaning.

Stuff which doesn't, uh, align with other datas that you'll report on. So the notion of audit trail, if you heard that term, but people with finance background definitely heard that term. Audit trail means that you can track every transaction across its journey. So to have like a complete audit trail and to be able to deal with data as an accountant are two major gaps that have not been solved in [00:10:00] finance.

So the more systems you have and the more automation you have, the more problems you have. Because first you didn't need to deal with more data. And second, you need to track more systems to be able to make sure the data is consistent. And these are the two major things that will obvious to me that in the era of ai, and we started just right, just right after AI started, but.

Well, obvious to me that, uh, uh, you needs to be solved and haven't been solved before. 

Mehmet: Now this is very enlightening, I would say. Like it's, it's very, uh, you know, well-defined problem statement. Ahikam and I'm happy. You mentioned something about like any organization with any size, they have to do this.

Like they, they have financial transactions, like whether, even if they are like a, a, a solo founder or whether they are like, uh, [00:11:00] hundreds of thousands of employees now. What can go wrong? So I understand the challenges and thank you for this great explanation. Now what are the consequences? So, because you know, when, as you said, excellent, 

Ahikam: excellent 

Mehmet: question because you said it's, it's, it's, uh, you know, we need to find the, the, you know, the, the audit trail.

What happens if we don't find it? I cam like, what are, what can, what can go wrong? 

Ahikam: Okay. So my, my most important thing is to be able to explain it very easily. So it's fun for people to. To listen and not like becoming like this, uh, complicated podcast. So think about, I wanna use an analogy. Think about you use your own credit card at home, or let's say you have a family, you have a spouse, you and your spouse, you're using your credit card.

So obviously I'd like to think that you use your card diligently, right? You, you only buy what you need or what you think you need. [00:12:00] Which are two major criteria. We buy what we need and we buy what we think we need. We may not need it, but we think we need. So you do that. So most people do that and apply judgment, of course.

Right? But most of us, what we are not doing is getting the statement every month of the credit card and reconciling it and checking that what we will charge for is actually what we bought. That's like the equivalent to governance in enterprise. Not only. That you allow people to act and spend money or whatever, you also need to check the data after they did it.

So not only that you have like signature right and approvals and all of that. This is all fine, but in the enterprise, unlike it, how fer, you also need to constantly check the data. Why do you need to check the data for multiple reasons. The first reason is. [00:13:00] Uh, for data integrity, meaning you wanna make sure that the right transactions are booked at the right place at the right time.

If you, um, had, uh, hired, um, uh, or did like a marketing activity and it was charged into r and d, then you have the wrong data, both in marketing and in r and d because marketing doesn't show the full experience. An L and d show an expense that doesn't relate, right? So you need to have the right data captured at the right time in the right place.

The other thing, so that's one thing, it's data integrity because your job as a finance organization to be the, the gatekeeper of the financial data, right? Using PCRM. The other thing is financial data is money. So if you don't monitor it, right, it's not just you will have the wrong sales data or the wrong marketing data, and maybe you'll approach the wrong customers [00:14:00] or whatever.

You would actually lose money if you don't track all your financial data. You don't track your revenues, you don't track your billing, you don't track your collection, you don't track your spend. You may end up spending, uh, paying twice for the same invoice. So making wrong. Um, actions on financial data means you lose money.

The third reason is compliance in the finance world, unlike the marketing world or the sales world or the engineering world, if you have the wrong data, it's a compliance issue. Whether because you are a public company and you need to keep your data in, uh, your, your data in order, or whether it's because for tax reason.

Think about that. For example, in the uh, uh, uh, uh, in the example I've used before, r and d, expense and marketing [00:15:00] expense may have different tax treatment in different regimes. Maybe you get like a special incentive for r and d or whatever. If you put the wrong expense in the wrong place, you are actually violating, may violate a law.

If you don't account for your VAP or GST or sales tax, you violate a law. So there are also compliance issues to maintaining the data. In many countries around the world, the chart of accounts, which is like how your books are organized, is dictated by law. So in finance, there's also a compliance aspect.

To managing the data. And last but not least, parts of your data are also customer facing. For example, you ask what can go wrong if you have the wrong billing data, because the way your CRM is connected to billing may have some bugs. [00:16:00] Or maybe in the contract you agreed to charge the customer for special uh, services or whatever.

And you didn't charge him. Comes end of the year. All of a sudden you charge him for everything at once. It creates relationship issues. It's customer facing. Think about it. It's not just your data and how you manage it. It's data that impacts third parties. So these are some of the reasons. Compliance, customer facing, uh, uh, data integrity and leakage.

Very good reasons why your data needs to be accurate at all times. Was that helpful? That's a 

Mehmet: very super, super helpful, and this is, you know, it showed the problem, it showed like the consequences or how the problem is and maybe you're seeing what I'm trying to, uh, to come from, so. Now what I'm curious about, you're solving this, uh, problem.

Uh, I am [00:17:00] using AI actually, so you, you, you build, you know, the, the technology utilizing ai and I'm sure there is a lot of information that even or small organization, they will have. Again, like when I was checking the website, like you have multiple use cases and I mean, the flow that you have put, you call it from the code to, to, to getting the cash.

Right. So from the moment I, I send a quotation to someone, uh, till the moment I get paid. And I, I believe maybe it's also like vice versa if I have to do payment for someone, correct me if I'm wrong. So how, what are you doing with AI here? Exactly how AI is. Helping organization, using your technology, of course, to streamline, you know, and solve these big challenges, which I'm sure the main guy who's thinking about it is A CFO, right?

So, so it's a big, a big headache for them. 

Ahikam: Yeah. So, uh, first of all, just, uh, touching your last comment [00:18:00] in, in large companies, in large companies, including in the Emirates, in the us, Canada, and in Europe, typically they have an organization or the CFO. He's the owner of the finance shop, but he has many, many roles including investor relations and talking to the board and you know, being kind of the foreign minister of the company going to, um, all kind of conferences.

So the CFO definitely is liable from a compliance from a. Account from an ownership standpoint, but typically it falls, uh, under the chief Accounting Officer and the chief accounting officer in the office of the CFO is responsible for the data integrity. He is responsible for, or at least ly responsible to better with it, and he's responsible for the compliance and all of that.

Uh, but of course it all comes down under the, the, the CFO [00:19:00] before. Answering your question, what AI is doing, let's talk about the problem for a second or why it wasn't solv today. 

Mehmet: Sure. 

Ahikam: The reason it wasn't solv today is in the finance world, you have like two, if I'll try to really simplify it. You have like two major.

Type of activities that are happening. You have systems, many systems. I would say 10, 15 systems. I can give examples. It's an ELP. It's a CRM. It's a CPQ. It's billing, it's spend management. It's HIS, it's payroll. It's um, uh, uh, uh, uh, all, you know, all kind of system procurement. Uh, where people are, uh, dealing with financial data and they transact on these systems, right?

And then at the end of the day, the job of accountants is to keep the data complete, to report the data, and they do all kind of activities. Uh, we call them like close financial, [00:20:00] close every month, every quarter, every year. And they're responsible for tax filing. So they're responsible for internal filing, external filing, all of that.

In the middle between the systems and all that work, the transactional work and all the works that takes the data and use it for reporting purposes, for filing purposes, for tax purposes, for all of that, you need, um, you have people in the loop and these people all, uh, and these people are responsible to, to check the data, govern the data, and they do a lot of manual activities.

Because these activities are complex, you have to have significant involvement of human in the loop, 

Mehmet: right? 

Ahikam: When, when AI emerged and now when ai, uh, you know, I'd like to think I use. I'm not an AI expert to be honest. I, I, you know, we all know how to [00:21:00] use, I think ma my job as an entrepreneur is to hire the experts to, but sure what I can learn from the great people and the great leaders in ai.

That, you know, you, you, you compare the maturity of AI today to a human being. So if three years ago AI was like a 3-year-old kid or five-year-old kid, you can talk to IMI Lance, he can apply. So thinking and then a year ago, AI was like a 17-year-old student. Now I think we are getting close to AI being like, uh, they compare it to like a, let's say 30 5-year-old PhD guy.

So in certain areas, AI really acts like an expert. Our ability today to use the existing AI technology, uh, uh, we leverage it in two or three ways. The first way is to understand the data of the company from all the systems, auto, uh, uh, auto, uh, uh, autonomously, [00:22:00] meaning mm-hmm. Think about, uh, a, a a a business transaction, a sales transaction in the CPQ system.

In the quote system, it's, it has its own data structure. In the CRM, it has its own data structure. In the, in the ERP, it's on its own data structure. The same, the same, uh, data point could be named differently in each of the systems because each of the systems will developed separately. So the ability of AI to take the data and understand the linkages between the data.

Because if you, to remind you we want to create that audit trail, we want to create that graph, that's one area where we can use ai. So AI can look at all the data and say, Hey, um, I think in the CRM. We call the date, transaction date, but the ERP we call it like sales date, but it's the same, it's the same date.

So I can take these two data points and compare and make sure they're all the same, so the AI can understand the data structure [00:23:00] from the different system. The other, the other way we are using AI in a unique manner is in finance, unlike many other, uh, areas in the organization. The source of tools is always a document.

When you transact with a company, you always have a document, whether it's an invoice, whether it's a quote, whether it's an order form, whether it's a contract, right? It's always a document. It's not like in marketing where you can create your own digital world and you have cookies, but it's not like a legal.

Finance, very much like legal. It's all based on documents. So you need the ability to read documents, understand them, and apply them to the data. That's the other thing that we are doing with ai. So today we can use agents for the first time in history to totally replace many, many activities and [00:24:00] chores, things that people do every time that.

You have to have humans think, for example, about a company that, um, uh, um, do a lot of, um, you know, they do a lot of, uh, transactions every month. Mm-hmm. Using some kind of a company. Every contract is unique. It has its own terms and exceptions, and someone needs to read and make sure to remind you. Part of the finance work is to make sure to govern the data, to make sure that the data was captured correctly, the starting date, the ending date term for termination, maybe special pricing.

So I charge you X, but if you go above a certain level, I tell you why. Someone needs to understand it and make sure it's okay. Now we can use a machine to do it. The machine can read if you, it'll take you, maybe, let's say, even if you are trained, your eye is trained. It'll take you five minutes to blouse on a five pages contract.

For the machine, it'll take 20 seconds or 10 [00:25:00] seconds. Pull all the relevant data and check it for the first time in history, we can use machines to compare between what we call structured data in the systems and unstructured data in the document. Why? Because every documents, people are getting creative.

They're adding clauses here and there and changing and all of that. We don't know it, but what we know, someone needs to read it and check it. We can do that with a machine today. So these are the two areas. The ability to understand the data from different systems and compare it and align it to what we call a financial data graph, which is our own proprietary technology.

And then the ability to also read unstructured data, documents, spreadsheets, whatever, and really do what people are doing.

Mehmet: That's fascinating. So Ahikam like, you know, Ahikam from a technology background and you know, in, in the IT department, we, we always [00:26:00] had this dream of having the real single pane of glass, right? That shows us everything in and out because, you know, like. In technology systems, of course. Like we have also this, uh, fragmented part, and I know I'm not a financial expert, but, you know, talking to to, to a lot of, uh, CFOs, accountants and, you know, uh, subject matter experts.

They tell me exactly, you know, what, what you just mentioned. So can we say finally now AI is bringing this single truth, you know, for us by whether making it work as an agent or whether I can chat with it and say, Hey, like I need to know. Where this is transaction, where is it exactly like why we so the way, 

Ahikam: right, right.

So we, we actually provide both, but it's more like the first example initially, how we engage with a customer. So, um, the way we engage with a customer like today, uh, for example, one example that comes to mind, we are working with a large, very large tech, tech [00:27:00] events company. They do corporate events using technology.

They're very large company, private equity owned. They have a large team of people that has to go through hundreds or even thousands of contracts every month because the company is very successful and they have a lot of document. And what we are telling them, instead of like having like 40 people doing it, you can have a machine and five people.

Why do you need five people? Because sometimes there are exceptions that you need to solve. Mm-hmm. So you, you have to have a human in the loop, but you don't have to add 10 people every. 20% of business goals you have, you can stop right there. Now. I'd like to think that, you know, a hundred years ago we used to use Abacus, you know, to calculate stuff.

Oh yeah. Now we use calculator. No one wants to go back. Right. You know, there are certain things that, you know, uh, you don't want to do like you used to do before, although, right. So I'd like to think that automation. As far as it [00:28:00] replaces some human in the loop. It doesn't replace all the human in the loop, but it allows you to scale because AI never sleeps.

It works on night and weekends and, and, and allows you to do, um, you know, no, none of us wants to go back to, uh, paying, you know, sometimes, you know, you know, paying digitally. You have your wallet stored in all kind of places and just click a button and you pay instead of typing 25 digits and names and, and security codes.

We like that. We like automation. So the ability now to use automation to do repetitive complex tasks in finance that help you save money, that helps you be compliant, that helps you keep the data in the right place. All of these things. I think it's like, um, something that was not enabled before three years ago.

It, it would've been a dream and now it's available and, uh, it's available in the areas where, again, you have to understand, you [00:29:00] know, autonomous driving is amazing. It's great. I think it's gonna save human lives, right? But, you know, people can drive today in finance. People cannot go through every transactions and do all the things that the machine is doing.

So today you are using sampling. They're doing whatever they can, but you can't hire people for every, to look at every single thing the machine can. So the machine can give you a much better compliance, a much better data coverage. You know, it can work tirelessly. It doesn't have a scale issues. So for the first time, we can move to a hundred percent check data in real time to make sure you are compliant, you are diligent, you are not losing money.

Your customers see the right data and so on and so forth. If it makes sense 

Mehmet: a hundred percent. Just quickly on this also, and you know, again, excuse my ignorance here a little bit. Ahikam. Uh, does that also, you know, [00:30:00] helpful in situations where we are, have some doubts? We are having some doubts about, like maybe, um, in, in some problems in payments for, which is mainly fraud prevention.

Like, can, can, how is it also helping in, in lab domain? I'll, 

Ahikam: I'll give you an example, a use case we're now doing now for. Okay, sure. One use case we'll do with two customers is something that can happen, you know, look, when you have human in the loop, people can make mistakes all the time. Absolutely.

Especially with like tiny little data, not too big decisions. We all make mistakes. It's human, but machine can reduce the level of mistakes that we do. So I'll give you two, I'll give you one example. Uh, two companies, uh, it's a very common use case that was not enabled until now, meaning it was enabled, but.

Manually think about the companies that, let's say, processes like a thousand invoices a month, payments. Mm-hmm. From, from, from [00:31:00] suppliers, from everything like office lease and insurance and, and this and that. Right now, it happens sometimes that the vendor changes his bank information. Right now, most companies today requires that the vendor will put his bank information on the invoice.

But the companies, when they pay, they pay according to what they have in the system. So what do they do? They have a manual process that tells the accountant, look before when you approve the invoice for payment check. That the data on the invoice, the bank data, the payment data for the vendor is the same data that we have in our system.

And if it's different, then let's flag and check out, right? Otherwise, you end up not paying your vendor on time. You pay and you can lose the money because you paid some other bank account and you'll have trouble with the bank retrieving the money.

This is a process that in, uh, two companies we're [00:32:00] working with consumes like a team of two or three people, and they also make mistakes because, you know, after like you see 500 invoices, you get tired. Really? Totally replace that with the automation. Wow. So the machine checks leads, the invoice understands that this section is about payment data.

Goes to the ERP system. Check the payment data there if it's okay. Clear if it's not okay. It flags wow. Totally waste of time for people to do it a minute after 2025. Um, but that's like a tiny use case. There are many, many use cases like that, but that's like a tiny use case that can save you time, money, um, you know, uh, uh, agony or, 

Mehmet: yeah.

Just, just one, one note on this I can, because you know, I've seen the other day, I can't remember where like some kind of debate about, especially, you know, these AI systems in finance. [00:33:00] Specifically in anything related to, uh, auditing or governance or like, uh, fraud prevention, anti-money laundering. And the thing is, you know, they were saying still, we need some, you know, at least a human to supervise this.

Of course. 

Ahikam: Right? So I'll give you an, I'll give you an example. I'll give you an example. It's, it's very important. A ML is very different. I actually come from the payment world, but I give you an example of what we mean, a human in the loop. So the way it works today. If a company has a thousand or 5,000, then typically they will have an offshore operation, maybe in India or the Philippines or whatever, where a lot of agents take this data.

Okay, that's fine. Right? And also these agents can make mistakes. So sometimes they can say something is okay, but it's not okay. But what do you really want? Where do you really want a human in dilute only in the places where there is a discrepancy Because what I, I promise you. If you go to a company that actually checks that 95% of the invoices [00:34:00] are okay, and maybe 5% are not okay, or maybe 2% are not okay, right?

But in order to find these 2%, you do the work for the other twin 98%. I used to have a friend lawyer who used to tell me, uh, it's a common joke in lawyers that 99% of the lawyers give a bad name for all the rest. So, uh, so, but that's not, so 95% of the invoices are okay. Only 5% are bad, but you need to check all the 95, right?

Because otherwise you wouldn't know the way to solve the discrepancy. You need a human in the loop to look at the 5%, but you don't need, you don't need all the resources to get to these 5% today, okay? So you can have the machine look at everything and when the machine has an issue. It could be sometimes a small false positive, as we call it, but when it has a discrepancy, you give it to the human envelope, but all of a sudden you need maybe 20% of the resources you needed before.

So instead of having a team of 20 people, [00:35:00] you need a team of two people that can handle discrepancies and the machine. And if you think about that, even in today's world, even when the, when the home computer was introduced 40 years ago, it helped us automate very small things, right? So that, that already happened in history many times over and over and over again.

What, where AI is different, that AI can get it to the ultimate level of, uh, of a human being, uh, uh, where you can, he can read, he can sing, he can compare, you know, before that most of the computer tasks were on simple automation. So even with that, I think when the, uh, personal computer was introduced. All the software allowed us to keep less people, but we still had to have many, many people to make decision.

Now agents can make decisions or at least simple decisions. So the idea is that you need a human in the loop, but you only already, you only need a human in the loop in [00:36:00] finance to solve the issues. You don't need to find out, and those are very common saying that says that finding the problem. 75% of the solution, right?

If someone tell you where the problem is, that's 75% of the solutions, right? If people don't feel well and they find the root cause, that's like almost the solution because then they can know who to go to to solve it. But a lot of times our struggle is where is the problem? 

Mehmet: A hundred percent. And, and thank you for, for clarifying this also.

I, I'm like, it's, it's a very important point. Now I want to go back, you know, you, because you mentioned about the role of the CFO actually, and um, and you repeated it multiple tasks, which also attracted my attention around the size of the team, right? So. How, how, how are you seeing, you know, the, the shift in the role of the CFO, [00:37:00] whether in small organizations or whether, like in the large enterprise organizations, um, how is it evolving and, um, you know, to your point, like I work in organizations before where the CFO was responsible for the IT department, for example.

So we, we know that for a fact that of course, like we somehow we were reporting to the finance, um. Some organization, they still separate them as A-C-I-O-C-F-O, you know, separate entities. But you know, to me, and I spoke to a lot of CFOs, at least here in the UA and, and some in the us they were telling me like the biggest thing.

Um, that now it's on top of mind. Keeping them awake at night is like, how us as CFOs, can we accelerate, you know, this transformation. It's not anymore about digital transformation, it's about also the AI transformation. What's your, what's your point of view on this Ike? 

Ahikam: I'll tell you this. Look, I think we'd like [00:38:00] to think that companies are made the same, but every company is different.

Mm-hmm. In every company, the head of IT responsible for different things and the CFO responsible. Some companies, the C ffo is overseeing HR and procurement, and it maybe sometimes it's not. It really comes down to the culture of the company, the culture of the CEO and the nature of their business. So in every company, the role of the CFO is different.

I'd like to think that in most companies, CFOs are the ones that keep everyone honest about their return on investment, right? Mm-hmm. What do I need to invest? What do I get out of it? Because otherwise you can have like sales will want all the money in the world. Marketing will have all the money in the will want all the money in the world, and so is engineering.

Right? And the job of the tough job of the CFO is to find, is to find the balance between all these and to find how I optimize the return investment in each of the areas. [00:39:00] I think this makes the CFO in my humble opinion, and if there is a, you know, if there is like a, you know, a ccio, then of course also the CIO in partnership.

Um, the most responsible guys to embed AI in their operations. So in order not to get confused, every company has two areas where they can embed ai. One is in their offering, in their product, right? If you sell whatever you sell, you can sell it with ai. Maybe if it's relevant. I'm not talking about the chair furniture, but, uh, you know, even if you sell.

Air conditioning today. Maybe you can sell like an AI features that adjusted to whatever, right? Right. So there's like AI in the product, which is one thing we're not dealing with. And there's the AI in your operations, how you make your marketing better or more efficient, how you make your, and I think the job of the CFO and sometimes together with the partnership with the [00:40:00] CIO depends on the culture of the company, is to become that chief transformation officer responsible to take the new capabilities in the world.

And make their operations more efficient and more cost effective and all of that. And I think that's the role of the CFO. And it's a tough, it's a tough job to understand what are the activities in marketing where you have to have a human in the loop, and what are these activities that allows automation to make you not only more efficient, cost effective, but even accelerate.

Right. So, and, and, and that's that that job remains the same, to keep everyone honest about what they need to do and what is the investment required for them to achieve it. So I actually think that as much as AI is being perceived by us as something which is like technology oriented, I actually think it's one of the major additional tools in the CFO tools box [00:41:00] to look to, to talk to the colleagues, peers in the company, in the enterprise, and tell them, show me how, at least we can try.

To embed ai. Maybe we'll try in five places and we'll get two winners, but we try and we see where it can make you more productive. And I thought, of course you need to all eat your own dog food, which is why I think we are trying to help CFOs demonstrate that they can easily and probably gain even more productivity and more, more efficiency and and less mistakes.

Doing it in the office of the CFO. 

Mehmet: Very, very, very, you know, again, 

Ahikam: with all new things you have to try. When you do marketing, you do AB testing. You don't know, you know, when you, when you a company and you would like to sell your product online, you try different messages and you just test it. You see what resonates.

Same [00:42:00] here. Every organization need to take like three to five AI initiatives. Some will stick, some will not. You have to try, but trying will get you there. 

Mehmet: Absolutely. Now, just because, you know, and I'm happy with how the, the, the conversation and the flow brought me to here. Um, because, um, like trying, you said trying, Ahikam and, um, trying without breaking things.

Now when we talk about finance and, and we, we explained this and you explained this, uh, of course, better than me. We are dealing with, you know, with kind of governance here we are dealing with some, uh, regulations sometimes. Uh, let, let, and sorry if it's a loaded questions because I know like a lot of guests, they tell me, you ask loaded questions.

I don't mean to, but. Why the regulations are not moving at the same, you know, pace as the technology is moving. And [00:43:00] do you think, or am I mistaken or do we need to, to put some more pressures to make sure that, you know. Of course we need to, to stay regulated. We need to stay, you know, in governance. But are, are there any efforts to, to also try to modernize these, um, you know, these areas as well, like how AI might shape even, you know, these, uh, you know, these, um, uh, finance, uh, governance rules and, uh, regulations down the road.

Of course I'm not, no one can predict the future. I know this, but again, from someone veteran in, in this domain like yourself, I can, 

Ahikam: I'll tell you this. Look, I think it's a very bold question, so I'll try to cut it in pieces. Yeah, please try to answer pieces because it's a very, very bold and I can explain why.

But, um, I'll tell you this. Um, first of all, president Trump in the United States couple of months ago offered that public companies will have to report twice a year instead of full times a year. [00:44:00] Which means like he sees regulation as something important, but you don't have to do it every three months.

So you do see different thinking in that area. But if I go back, I would say that in the era of AI regulations become much, much easier to meet because a lot of the regulation. We know required a lot of manual work, a lot of documents preparation, a lot of, uh, data capturing, and think about how AI can behave like an artificial employee and do all of that in minutes in seconds, at least cut some of the work.

So again, if you think about the regulation in finance today, it doesn't require you to check every transaction. No finance. Every finance regulation assumes you're gonna make mistakes. It doesn't ask you to be a hundred percent accurate, it just asking you to be materially accurate. But [00:45:00] now with AI, you can be a hundred percent accurate, so you can actually exceed regulation, because at the end of the day, we all wanna be accurate.

It's just a question of what's the price of it and what's the ramification, what's the penalty? So I'd like to think with AI today. Your ability to stay compliant and meet regulation makes it way, way easier in many domains, in many, many domains because, you know, um, because AI can help you process a lot of information much faster, and it can understand exceptions, it can identify anomalies, it can help you prepare documents and reports and things like that.

It's, it's actually, I, I, uh, I, I, I think it's like, uh, for many, many reasons, regulation, it's, it's, uh, becomes, uh, um, I, I actually think that the concern, if you heard in the industry is that regulation [00:46:00] will not meet the capabilities of ai definitely when it comes to what can AI can produce. So I'd like to think, at least in the co world, which I'm familiar with in many areas of the co world.

Your ability to stay compliant with ai, uh, that can behave like an employee that never goes to sleep, sees all the data process, everything in real time goes much better in terms of like, things like data privacy and vulnera, security vulnerabilities and financial vulnerabilities and fraud, right? Because before that, to meet regulations, it's always, you had a very simple price, people.

People that needs to maintain policies and people that needs to do attestation, data attestations, and people who needs to do, uh, compliance reporting. All of that now could be, so I'd like to think as far as regulation con are concerned, um, uh, I think we should expect that it's easier to meet Reg, it's [00:47:00] easier to meet and exceed regulation requirements if people really want it.

Mehmet: Yeah. And thank you for, you know, decomposing my questions. This is exactly what I was aiming to, to hear from you because, you know, when I'm not that expert in the domain, I try my best to put the ideas, uh, in the best way. But thank you for the expansion. I can, you know, final question before we, we, we wrap it up.

Uh, I can this, and this is for founders, um, who are currently building or willing to build in, in. The finance, uh, vertical because you've worked on both sides, like you've been in, in, in startups, you've been in corporate, and now again, it's, it's your startup. Final maybe words of wisdom for, for founders in this domain.

Ahikam: I'd like to think that, uh, finding a pain, uh, the, what we call the product market fit is the most important [00:48:00] thing to focus on. When you try to build a new venture, the ability to identify a problem and be able to explain yourself before you explain to investors why do, why you are the most suited person, not only to address this problem, but to solve it, and what are you going to solve it with?

And I think being able to understand, identify a pain. Then understand why, how to solve it, and to understand what's the business model around this. This, the solution are two, three major, major, major components that are extremely essential in building a new venture. Think about if any of the listeners had a chance to build a house or an apartment, right?

There's never enough planning you can do before they actually start to build, right? Because the more planning you do on how [00:49:00] to design, um, you know, um, um, uh, a place, whether it's a. Residential or business or an office or whatever, the more planning you invest in, how to locate things, where to locate things, lighting, uh, air conditioning, uh, uh, uh, you know, uh, storage cabinet.

The more planning you invest, the more mistakes you're gonna avoid. And the same applies for new ventures. The more planning you do around what problems I'm, I'm looking to solve, why am I the right person or why my solution is the best to solve? And how I can create a business model around that. You can't, you can't, um, uh, the more time you invest in that, the more successful I humbly think you're gonna be, um, uh, in that new venture.

Mehmet: Great. And what a great, uh, advice especially about, you know, the product market fit, the, the problem and, and, you know, finding, and this is why exactly I was, I was asking the questions [00:50:00] in the way I was doing also as well. Um, uh, I think I can, people can find more about Safebooks on the website, which is Safebooks, uh, dot ai, right?

Ahikam: Yeah, which we change every week. Every week we add more content. Every week we add more. So every week you change the, if you, even if you've been on the website last week, next week, it's gonna be different. 

Mehmet: Great.