Ahikam Kaufman, Safebooks | theCUBE + NYSE Wired: Business Transformation Edge

Palo Alto studio connection Silicon Valley and Wall Street. I'm John F co here with Dave Volante my co-host. Hello, I'm John Furry your host of the cube here in our New York Stock Exchange Cube studios. Of course we have our Palo Alto studio connecting Silicon Valley to Wall Street. This is our transformation edge series where we talk to leaders who are building the technology, operating the tech, investing in the technology to bring in this next era of AI. When injected into a company, intelligence, AI intelligence transforms the business, the business model and all of the operations. As a result, the entire business changes. That's the topic we're exploring. Akam Kaufman's here, co-founder and CEO of Safebooks AI entrepreneur taking a large scale approach, a systems approach to AI in the financial area, but of course has implications everywhere. Good to see you. Thanks for coming on the cube. >> Thank you, John, for having me. I'm excited for to to the conversation. >> So, finance has been this cottage industry. you know they had applications a lot of vendors would sell stuff to finance departments but now with the AI and postcloud era now we're in the AI era which builds on top of cloud largecale finance capabilities can be done fairly easily and with Agentic and AI injecting intelligence is certainly going to impact the overall company's numbers and finance actually likes to roll those up so they're going to be involved they have their own department needs so you see finance in the middle of all action. In fact, just this week, Jensen Wong was on TV and I wrote a post in April. He always says energy bounds AI. Well, I wrote a post in in April that said finance B is a bounding function. They announced Nvidia with Goldman, KKR, Brookfield, all the top capital 500 billion dollars for AI capital. Finance and energy are the bounding functions of AI. But so the word finance is in there. you're going after this market, right? So, we're we're going after corporate finance. And I'd like to think three, four years, three and a half years after the models came out, it's now finance are being asked, what are you doing about AI? The challenges or the problem was always was that finance unlike other parts of the organization they require data integrity and accuracy like no one else because they're subject to compliance and and you just need to rely on data which is accurate. So we created technology that uniquely provide AI with the right context and the right um um guard rails to make sure that every response would be accurate and agents would be able to operate as if not better than human beings. So um you had a previous venture you sold to QuickBooks I'm sorry into it right not QuickBooks is the product Safebooks QuickBooks any you get the green light to go with Safe Books on this one. I wanted to correspond with uh the the you know the QuickBooks uh brand. H but I wanted also to pick a name where it kind of like reflect the the when when customers buy a product I think they they they you need to kind of sell like um a sentiment. You need to sell an experience and making people feel safe about the data because every large company you there is explosion of data and you can't check everything all the time. We're creating that cyberlike experience for finance because with the automation comes also the level of accuracy and integrity that human beings cannot do uh uh with that level of data. So creating that sense of safety and security for financial data for the financial leaders is what our name kind of >> yeah and you mentioned accuracy um because of compliance but also they're in the numbers game they want to be they have to be right >> right >> because if you get a number wrong think about the consequences on audits >> you got to unwind it you got to throw human more forensic work at it so I think that's a key point but I want to ask you about your idea around this venture because in hearing you talk and I'd love you expand on this because I think you'd illuminate it feels like a platform because you're talking about >> right >> not just a point solution. >> No, >> you're you're talking about a platform. Take me through your thinking around this and what it means >> to the company not just the finance department. >> Right. So from the get-go we we understood we needed to do two things which are different. first to develop the technologies that can provide AI with the right context and again the guard rails to be able to be uh uh to be able to kind of like uh have like a deterministic way to to uh to make the actions. the the other things that the other things that we wanted to do is we wanted to to make sure that we don't just generate a single point solution but eventually with AI because you can do so many things I don't want the office of the CFO to deal with multiple or they don't want to deal with multiple vendors each for a point solution once you we've got the data right from all the systems we can using our aentic layer we can create that point solution on top of the platform allowing finance to go and expand their use using the same platform the same platform but instead of using the clouds of the world which are great but cannot provide that level of data accuracy and integrity they're using safe books and using our aentic layer you can configure any agent to do any work whether it's across cash revenues expenses and so on a data layer in finance is very compelling to me I like this idea So explain more because there's technology involved. Um we're hearing more and more um words like ontology, graph databases. I mean Palunteer I mean they popularized the term that's been around since the 80s. I did I did that in the in in as a kid. I was playing with ontology. Not a lot of infrastructure support the scale. But now we have a database world where you can actually connect data. You can put a context graph together. And there's a lot of benefits that are is compatible with computer science of AI recursion things like that are like you're starting to see a formation of a brain >> right >> metaphor why wouldn't a company want to have a financial brain and connect it to all aspects of the business >> I think again the technology did not exist even portion of our technology did not exist even a year ago I think 2026 is the year where finance are starting to embrace AR and embark working on that journey. I'd like to think that today 100% of the companies uh meaning finance organization in companies are being asked what are you doing on AI? In 2024 2025 it was more like engineering, sales, marketing. Now it's finance because we do have the technology to be able to provide accurate results which are also auditable. We recently acquired our sock one certification which allows our customers and their auditors to rely on the outcome. But the the foundation is the ability or the the the the pain point originally was the fact that in corporate finance every transaction it's processed by multiple systems and the ability to connect the dots between these systems was manual work and now it's fully automated using our >> Yeah. And you have connective tissue with with APIs and MCP servers. But you brought up a good point because compliance finance has all the elements of success built into for AI to to be injected. It's got a lot of stuff that's in manuals process rules. It's got um data probably lab good data not maybe not perfect clean data. Um it's got process and if you connect to other systems in a distributed computing fashion you now have a mechanism to cross boundaries outside the department. Hence other business units, >> right? I'll tell you even more. I think uh I think an example maybe Elon Musk used was that 100 years ago people used to sit in a bank and calculate the balance for each account on a monthly basis. There's works that I don't think anyone wants to go back and do that. I think we're going to witness that same trans transformation in corporate finance. Imagine people are doing the same work every day across the data. You can't do it across all the data because we're all humane and again there's a lot of details. So the ability to use a machine that can that can emulate or replace some of that manual work allowing people to focus on more strategic is actually essential but technology did not exist. I think a year a year from now it's going to look totally different. Well, the thing about that metaphor of counting money, but take the modern version of that preai labor is a essentially a manufacturing operations concept. It's an assembly line of people. >> Exactly. >> Who is spec to do this repetitive [snorts] task, the next person does the same. So you have uniform deterministic tasks, >> right? >> That's human. That's easy to understand. Just just to give you an example, imagine today in large companies, let's say uh they process thousands of orders for their customers every every single month. Now, finance needs to sit down, look at the documentation, look at the policies that they have, look at the billing, and make sure everything is correct and to approve the deal to this. And it takes them time until they approve the deal. The machine can do all of that which means that less mistakes, less customerf facing mistakes, less leakage or missing bills, billing or whatever. Uh better relationship with your sales organization because you move things faster and the anomalies and discrepancies are still so you manage the exception. You don't need need to manage the standard. >> You know, Aam, that's a really good point. You brought up something that made me think and I want to bring this back up. You mentioned sales and finance. In every company that I've worked in, yeah, in every company my friends worked in, every company that I've observed, >> they hate each other basically. And you got to put the finance department in one building and you put sales because sales are like they're very agile, the front of the office. Um, they think differently. the cultural intersection what you're getting at is actually a a a liberating opportunity from a change management perspective because finance people want things buttoned up and say ah just get the quote out there and then put it in that bucket so there's a lot of inbound logistics >> if you have a problem I want to know about it fast don't drag me for a week and then so speed velocity here is an essential component in the relationship >> how do you see finance with AI the way you're thinking about it impacting the veloc velocity, accuracy, efficacy of sales and marketing, specifically sales, go to market because product market fist change, but like you said, they're just on the front lines creating marketing, the value proposition of a company and then finance has to inside out align with that group. They're customer focus with a lot of potential error opportunities there. How does this fix that? So I'd like to think that you know I think eventually everyone needs to be a corporate citizen and everyone wants to to to get uh the deal done uh for the benefit of the customer for the benefit of the vendor the company and to get their commission check right and in order to do that uh they they expect their finance peers to uh to be able to look at stuff in time so they can react to the customer and be able to identify all the issues in details so if there is an issue they don't miss it and then something happens down the road. So the ability to accurately scan everything 100% in real time. So the velocity and quality of deal review is one of the things one of the many things Safebook solves. and the game. >> That's a huge value proposition because that's where miscommunication, misinterpretation, lack of understanding the other party, that's where this could be very harmonious. >> Right. Right. So, and again allowing people to focus on managing exceptions as opposed to managing >> everything. Yeah. We all seen deal reviews. Okay. Yes. And then send an email. It's asynchronous work. It's just that work gets goes away, right? >> All right. Talk about the company. Where are you guys now? take us through the the plan, the strategy, where you are on momentum. >> I love the platform play you have here. Yeah. So basically uh you know um we started the company three years ago uh we've built a very deep technology and also when the recent models came out around like maybe 9 months ago we were able to then uh amplify it with the agentic layer and we started to go to market. Now we serve about like uh I think 15 uh 15 large enterprises corporates and um basically we uh we will be starting to think about our series A soon and uh that's it the the >> how many customers do you have now? >> So 15 are they are they pro um are they design partners early customers? Oh, they're paying customers. Okay. Paying customers. We we provide value. Actually today the way the relationship go we prefer >> to do like uh to start the relationship kick off with a proof of value and we turn this around. We turn the use case around in like couple of weeks >> and then the interaction is different because we show you stuff on your data. It's not like >> it's like a best forward deployed engineer motion. You go in, look at their data, and come in with not just a slick demo, right? >> You actually have actionable. >> And then we want to become your trusted advisor because we know your data as well as you do. And with our platform, we actually would like to continue and deepen the relationship by expanding. >> Okay, you got me at at at a yes. You got me to yes. I'm a customer now. I want to work with you. I want to sign on the line that's dotted. Tell me what's next. How do I work with you? What do I need to do on my end? Right? How do you add value? >> Typically, our initial goal is to automate your closed processes. So, we again we have a four uh a team of four deployed engineers as part of our delivery team. They would go and map your closed processes. When you when you try to nail it down, it could be anywhere between 150 to 800 tasks depends on the size and complexity of the business. And then we go ahead and configure them on the system. It could be a process that would take depends on the size of the company between anywhere between 8 to 10 weeks to maybe even six months. But every single week or months, your life gets better and something else goes away. And when we create these agents, they can do three things. They automate the work. They create the documentation, the workpapers and all that specific documents, the proof of work. That's great, >> right? The proof of work which is essential. And you can interact with them. You can ask them questions like like a staff member, like an employee. You can ask them questions. They can help you. >> Those questions could be like, "Hey, why didn't why didn't you do this there?" Or imagine like if we >> can you get down to the level of why did you book it there in that category versus that, >> right? The same way because look in finance, it's critical especially in public companies that there will be a prepare and a review. We replace the prepare. The reviewer is still the human in the loop. The same way the human in the loop would ask his team >> what percentage of work would you say goes into preparation without AI. >> So I'd like to think it's like 90%. 90 heavy lift >> let's say. Yeah probably people you know 8020 but I think it's really 9010. >> I've heard similar. Yeah right. >> That's consistent what I've been hearing. >> Right. So it's like >> that's a lot. >> Right. That's a lot. It's much easy to review and also you think more strategically when you need to prepare every line. It's way more difficult and like you get lost in the woods as opposed to uh >> Yeah. It's the bark of the tree. Everyone's looking and they not see the big picture. >> Right. Right. So being a being able being able to do that and then you interact with the agent just like an employee. You can ask him questions. Uh you you you can you can challenge him. You can alter the documentation. You can >> All right. What's next for you? Tell us what's on your agenda. Obviously, you got 15 customers you're building out. You're getting more data, proof points from customers. I'm assuming to continue to get more customers. What else you got going on? >> Uh, so we're looking to raise our series A by early next year and uh that's really where we want to go for and and continue from there. I think our technology is unique. I think 100% of finance teams around the world around the country are looking at AI >> and uh I think it's the right moment to expand. >> All right. What would you say if I said well there's a bubble going on? How would you answer that question? Everyone's saying there's a bubble in AI. Um oh supply is constrained. There's demand for intellect intelligence. I think like any anything else in the world, the cost of using AI will drop will continue to drop down as the infrastructure will be more efficient and I'd like to think a AI would be even avail you know one of the analogies I'm using I think we're looking at the 40 now in 1920 and we're switching we're moving from the horse to 40 but I think the Lexus and the Mercedes-Benz and the Tesla are like 3 four years down the So I think we we we're not, you know, it's it's just getting started. >> There's a discovery journey for enterprises right now. They're learning. They're and once they get unit economics nailed down, once they see that intelligence, >> that's the product they're buying. >> And the world is going to significantly change in ways we can't even imagine. >> Yeah. Yeah. I think the demand curve is good. Now on the funding, have you had a seed round? >> So we raised a slight a relatively big seed of $15 million. >> Seed round. >> Seed round. It used to be the old series A, but >> no, no, no. $50 million of seed, uh, four or five different funds. One of the leaders is Propel Ventures from San Francisco and Inside Partners and some others attendee and um, and u, and that was really meant to help us build the foundation, the technology, and to support grow. >> So, you're going to have a good A round because you have some nice investors, world-class investors in there. Insight, >> very strong. Um, you probably don't need any help from me, but if you do rent from the cube, you get a 20% discount on the equity. >> There you go. >> He's he's [laughter] like, "Yeah, approved by the board." No, I'm only kidding. Good luck on the financing. Love the venture. Again, I love the systems approach. It's very consistent with what we're seeing on AI infrastructure. Kind of the same game, but everything's denser, close to the compute. You're seeing >> the architecture is systematic. It's a systems game. Congratulations. I'd like to think the world is changing, but we're also solving pains that people are experiencing on a daily basis. And once we solve that, I don't think anyone would like to >> to I mean, prep, it's just grind. Take that toil, that undifferiated heavy lifting >> and move it into review and >> whatever professional service they can do. That's the winning hand. Yeah. If you can get it done. So, congratulations. Thanks for coming on. >> Thank you, John, for >> All right. This is the kind of conversation at the cube we love. It's deep tech meets reality. That is the transformation edge. The competitive edge for the companies going forward will be how they can leverage AI. Not just leveraging for the leverage AI's sake. It's injecting intelligence into their organization. It's not an IT project. The payback is there and there's money involved. It is a transformation time and the best edge will go to the winners. Thanks for watching.