← SaaS Interviews with CEOs, Startups, Founders
SaaS Interviews with CEOs, Startups, Founders
Selling Check for $400M, Now Building a $1.5M ARR AI Startup
00:24:18
Loading summary
Transcript73 lines
- [00:00]
A
I think that was a $360 million acquisition by Intuit in 2019. Is my timeline, right?
- [00:05]
B
Actually it was close to $400 million.
- [00:08]
A
How many millionaires did you make?
- [00:09]
B
At least 10.
- [00:10]
A
Are you comfortable sharing what's the largest company pay you? Do you have any million dollar per year accounts?
- [00:15]
B
I would say our largest engagement is around $300,000 right now in Q2 of 2014 when we sold the company to Intuit, we were like the largest M and A deal according to the Wall Street Journal. By coincidence.
- [00:27]
A
Are you comfortable sharing what you personally took home when you exited into it?
- [00:30]
B
Very loud to audience. So I would prefer not to share that if that's okay.
- [00:33]
A
That's totally okay. How many paying customers are you working with now today at SafeBooks?
- [00:37]
B
So we have about 15 paying customers. You have to understand we are selling powerful data and automation platform for the office of the CFO.
- [00:45]
A
You got 3 million bucks of extra moving around. I want to invest. Hey folks. My guest today is Ahikam Kaufman. He is a veteran fintech executive who previously co founded Cheque, which was acquired by Intuit and served an executive at HP and Mercury Interactive. Today he's building SafeBooks AI which uses Agentic revenue integrity to autonomously ensure financial data accuracy for enterprises. I come. You ready to take us to the top?
- [01:10]
B
Yeah, hopefully.
- [01:11]
A
All right, that is a mouthful, but it's important work. Give it to me like a kindergartner. What are you selling today?
- [01:16]
B
What we are basically selling is the ability to understand your financial data across system, validate it, check it and replace significant manual work which today is being done by accountants. One of the challenges in the office of the CFO is there is a gap between what an accountant is trained to do and what he needs to do, which is basically check his data across multiple systems in real time across structured and unstructured data. And we now using the benefits and power of AI, we can fully automate that work.
- [01:46]
A
Let me give. Let's play a game here. I'm going to do old and you're going to tell me the new way to do it using your technology. Okay. Old way. I do invoices via bill.com, i have a fractional CFO. I pay 3,000 bucks a month. They go into my QuickBooks at the end of each month. They click, close everything out. They integrate Bill, they integrate my bank accounts, you know, via Plaid or Soft Edge or Teller. They close the books. That's the old way. The new way is what?
- [02:07]
B
Sorry, so that's like a Great example, but doesn't fit our icp. We're catering to large enterprises, companies, public companies, who need to execute governance across their data. For the most part they're not using the platform you mentioned and they're using SOF other systems. And the way they do business and conduct the business where we currently focus, which is all order to cash and revenue. And billing integrity is very, very cumbersome. Manual and process through multiple systems like CRM, billing coding, ERP and so on.
- [02:39]
A
So I did it on purpose, right? Can you give me an example? Just like I did. So instead of using CRM or cpq, actually tell the story of the actual, like an actual example of what you're replacing, right?
- [02:49]
B
So think about a company issuing a quote for a product or service, right? Which then translates into a contract. The contract has many terms, it's complex, but the contract is not fully accurately captured by the CRM, which may lead into, leads into billing discrepancies or errors. And then there's a human in the loop in finance that manually checks the data across all these systems and across the document, right? That happens when you close the deal. Then maybe you have to charge the customer extra for all kind of extra usage so he goes back to the document, he changes stuff. Then maybe the sales organization, they changed some of the terms with the customer. So again the data is wrong. And today all of that work has to be done by people where they are logging into the disparate systems, checking the data, checking the documents, which is always the source of truth. And we now can replace it with agents who see the data end to end and it has to be end to end. And that's part of our secret sauce. And when they see the data end to end and they understand it, they know exactly where the discrepancies are and how to remediate them. So.
- [03:57]
A
So this sounds to me more like it is. You know, there's an industry that's well known called quote to cash. You're effectively making that system way more efficient using artificial intelligence and agents.
- [04:07]
B
Yes, we are focusing on billing and revenue integrity, but almost the same use cases exist in procure to pay and in payroll. These are like the, the main three engines that leads most of the company's money flow or cash flow.
- [04:22]
A
So yeah, are these companies that need at least, you know, 100 million of revenue before they feel this problem and pay you, or is it bigger, 500 million revenue or more?
- [04:28]
B
Totally. When I'd like to think that companies that starts to exceed like 2,300 million dollars in revenues would significantly feel the pain because they would have to cope with multiple offerings, multiple products, data structure which is very different between one system to another and a lot of transactional volume in their day to day business. So on a monthly, quarterly, actually on a daily, weekly, monthly quarterly basis, they have to check, verify their revenue data. The reason they need to do it is for three or four very critical reason. Customer facing data, right. If you're wrong in your billing, it's not a pleasant experience with your customer. You need to check it for compliance purposes. Then you have to deploy people. Now the industry is facing an accountant shortage. It's not happening in real time. People can make mistake. Even if you offshore this work, we see significant amount of mistakes in each company we're working with. And it's all natural in you.
- [05:24]
A
I think the product suite now is super clear. Thanks for that. When a customer does sign up and use you on average, what are they paying you per month or per year to use the technology?
- [05:33]
B
Like to think that the initial use case we start with, it's around the way we kind of sell it. It's the cost of a single resource, finance resource. So it's, I would say it's around like 100, $125,000 for the initial starter use case. And then each use case has its own roi. So you can justify that and charge more. But what we are now doing, which is pretty amazing, we are now releasing a capability which allow the customer using AI to prompt and configure his own use. The more work the customer can do, it can do on the platform serving their customers, the more they can benefit from the system without adding additional cost. Because we don't sell basically adding data.
- [06:16]
A
My gut tells me because you're adding so much extra product value that you have upsold customers way above the $125,000 ACV without naming obviously the customer name for confidentiality reasons. Are you comfortable sharing? What's the largest company pay you? Do you have any million dollar per year accounts?
- [06:30]
B
You know we just started to go to market like about less than a year ago, so. But we do have, I would say our largest engagement is around $300,000 right now. But we do see the potential because they continue again companies continue to suffer. We are now seeing a significant reduction in stock prices which will force companies to be more efficient and try to remove the human in the loop as much as possible.
- [06:56]
A
So I think just to be clear on the growth story, it sounds like 2025 was your first paying customer. When did you write the first line of code for the platform, what year?
- [07:04]
B
So we actually started first line of code probably around June 2023. But the thing is, is that in order to do what we are doing today, we actually started from the foundation. And the foundation was to build a sophisticated proprietary graph database that for the first time connects and links automatically using AI all the various data sources in the office of the cfo. That was never done before. And that allows us to easily automate because we provide the agents the end to end view on how the transaction behaves across systems and linking it to the source of tools, which is always a doc document.
- [07:41]
A
How did you fund the business between 2023 and 2025? I know you had obviously a big exit in the past. Did you self fund it or did you, you know, raise capital?
- [07:48]
B
We raised a $50 million seed in like two chunks from like six, seven funds. Early, very early stage fund, but we raised like $50 million.
- [07:58]
A
Got it. That makes a lot of sense. Guys, remember, I am not just a YouTuber. I'm investing in my third fund. We've deployed $250 million into 550 software companies so far. Again, at founder founderpath.com if you're interested in capital, I would love to cut you a check because I know you're investing in your education. You watch my show. So sign up@founderpath.com and when you get the onboarding email, I reply and I see all those. Just reply and say, nathan, I found you through YouTube and I'll make sure to prioritize you. I would love to cut you a check. Check out founderpath.com we got to get your backstory here because you're an impressive founder. I come. I mean you've done this before, going all the way back to 2014 and scan modal. But I think the big one that most people will know you by will be the check story. I think that was a 360 million doll acquisition by Intuit in 2019. Is my timeline right?
- [08:44]
B
Actually it was close to $400 million, but yeah, that was. And funny enough, in Q2 of 2014 when we sold the company to Intuit, we were like the largest M and A deal according to the Wall Street Journal, by coincidence. But at that time, these type of deals were pretty big or considered pretty big.
- [09:02]
A
You spent a big chunk of your life on this. Launched in 2007. You're, you're in it for 12, 13 years. Give me some context on that business. I Did you raise capital there as well or did you Bootstrap.
- [09:11]
B
So you know at we were trying to actually it's interesting to share. So at Check we wanted to help people organize their personal data. Right. And we started that in 2007 and then we realized we decided to focus on finance personal financial management data. And in order to do personal financial management data at the time cloud did not exist. So we decided to create our own data foundation and our own links to the banks. In order to be in a position where whatever product we build on top of it, it will lay on a very structured sett set of data.
- [09:45]
A
We If I'm just clear on that, just to make an analogy, this was before Plaid and Salt Edge and Teller these IPASS tools.
- [09:52]
B
Totally. You had a company then called the Yodlee and maybe another company and. But we decided to create our own data foundation which is by the way was adopted by Intuit and using into that same technology is using Intuit is using today for their bank fee. So we decided to create our own and that was a tough decision. Why? Because we knew that consumers eventually will have to rely on our data for their personal finance management decisions. And that data has to be super reliable. At that time we had a competitor called Mint which was a very popular mobile application. But they bought their data so they relied on a third party data infrastructure again before the plot days which was less reliable, causing a lot of mistakes. But so we spent like a year in the garage building our data foundation and only once we did that, we started to build the application on top that would serve you with the value you need. Like you know, seeing all your data from all your banks and credit cards and being able then to see all kind of anomalies in your data or fees or something like that. And then on top of that we build a payment platform allowing you to pay your bills from these connections. Because we own the connection or the bank feed. We could do like a two way interaction with the bank or the credit card and actually use that use that connection in order to trigger the payment. So the investment in data which caused us to launch the product later than our competitors but launch more robust product. That what also inspired us when we when we started to work on safebooks. Because the idea is that you can reliably bring or create a financial data lake which creates like kind of a unified data set for all the financial systems and link the data together. Because in finance it's all about the linking. You want to make sure that the quote is connected to the opportunity in Salesforce which connects to the right set of invoices or billing, how you build the customer. And only when you do that you can actually ask the data questions and check the data. So that's what we did back the day at In Czech. And yeah, that's what works for us very well.
- [11:58]
A
I believe you chose to go down the venture route, right? I think you raised about 50 million at that business before exiting, is that right?
- [12:04]
B
So, yeah. So to your point, we raised the initial seed in 2007 from a group of private investors and then we raised our A from BC called Pic, and we raised our B from a fund called Morgan Thaler, now it's called Canva. And we raised our series C from Nello Ventures. Doug Carlisle at the time was a partner. He retired since then and he decided to fund us and gave us $25 million in December of 2013 and we paid it back in June of 2014.
- [12:36]
A
Good return for. You're his best friend now.
- [12:38]
B
I imagine he, he retired since then. But I think I, I put in my LinkedIn, they published a post this was like one of their highest irr because you know when you return, when you do like 3 or 4x, which that's what they did over the course of like six to eight months. It's like, yeah.
- [12:54]
A
How did you personally manage your own dilution at that business? Are you comfortable sharing how much you owned before you exited it?
- [13:00]
B
I think we were two founders and each of us had like, you know, between 5 to 10% of the company. You know, I think our employees also did very well. And you know, I think it's all about how do you distribute the equity with your employees. And by the way, one of the things you learn when you sell a company is that at some point in time it becomes, comes too late. You can't reverse these decisions. So you have from the get go, you have to decide how you're going to compensate in equity. Your founding team, your staff member, so when the time comes they, they benefit from the results, you can't repair it afterwards. So we had that and again we, we raised like $50 million, which back then was a lot of money. So we got diluted. But that's fine because I, we, you know, you really, what you're really passionate about is building, working with a great team that you need obviously to fund. And the, the, the more team members you add at that time, obviously you couldn't do vibe coding. They had to hire people to write code. So the more, the more team you have, it's, it's easier to get things Done for the most part and obviously also to market the platform and so on and so forth.
- [14:07]
A
I come, I'm asking these next couple of questions not because I, you know, I won't necessarily brag about it or I'm interested in your personal finances, but I think what you did is something that other founders should aspire to in terms of, you know, not always raising at the highest valuation, managing dilution, taking care of your team and employees. At the end, are you comfortable sharing at, at exit? How many millionaires did you make?
- [14:26]
B
I'd like to, you know, we, we were not like a super large team or probably about 80 people at the time, but I'd like to think at least 10. At least 10 like the founding team members became, it was like a life, definitely a life changing event. By the way, one, one thing to keep in mind when you sell a company, smart buyers and Intuit is definitely a smart buyer and I have all the appreciation in the world for this company for many, many reasons. And you know, the way they build products and the way they execute is second to none in the B2C and the B2SMB market. But the acquirer will appropriate, let's say 90% of the fund to the shareholders of the company, founders and employees included, and then typically 10% for retention. And we also took that retention chart which at the time was an additional $25 million and we fully deployed it across the board board to people telling them the following, listen, we can actually, we have an opportunity now to remediate you. So even if you are not part of the founding team, but you've done a great job over the last year because you only joined a year ago, we give you that money, that allocation of the retention because of the future work you're going to do for the combined business. So we were able to actually make additional team members very positively impacted by the transaction. You know, allowing them to focus on not worrying about the future, to focus about the integration, focus about the new products and offerings that we're going to create. And that worked very well for us. Actually most of them have stayed until this very day. That's amazing to hear.
- [16:04]
A
I love hearing that. And look again, you don't have to answer this, it's a personal question, but I do think it helps founders understand when there is a company that sells for 360 million and they read that in a headline today and the company's raised 50 million, you're always wondering, I wonder what the founder made. Are you comfortable sharing what you personally took home? When you exited into it.
- [16:19]
B
I actually you have a very large audience, so I would prefer not to share that, if that's okay.
- [16:26]
A
That's totally okay. How do you know I have a large audience? You listen to a couple.
- [16:30]
B
I listen to many of your podcasts, but I think, you know, I know your content reaches a lot of people. I heard it from many people.
- [16:38]
A
So yeah, that's fair enough. I won't push you.
- [16:40]
B
I think your passion about SaaS and hopefully now AI is like, I've created a dent in the, in the marketplace.
- [16:48]
A
Well, that's very kind of you. I think you're probably giving me too much credit, but the reason I bring all that up is you were able to raise that round for your new company in 2023 because you created a great outcome for investors. You're a proven founder, you're a proven entreprene. You've now closed that seed round. Total 15 million bucks. You mentioned your first paying customer was in 2025. How many paying customers are you working with now today at SafeBooks?
- [17:10]
B
So we have about 15 paying customers. You have to understand that in the. We'll look, we're selling, I think, a super powerful automate data and automation platform for the office of the cfo. And sometimes again, each company has its own processes to make sure that the A doesn't like, go crazy. Right. Although I think what we do in AI is very, very different than the models out there. We don't have any hallucinations. Our accuracy rate is 98%. We actually develop tools allowing our customers to run their own UATs to understand the level of qualities that we provide. But I come.
- [17:48]
A
How are you using hallucinations? Are you setting up a private RAG database for each, like a data lake for each customer that signs up with you and are just talking to that ragdb to hallucinations?
- [17:57]
B
That's exactly what we do. And on top of that, we augment the results with rules that allow us to make sure that we understand that the results are on track. And if something is not on track, because all the enterprise data is very structured, I can understand the outcome based on the various data points I'm looking at. So I can understand, for example, that let's say I'm pulling a data point from a contract and I now compare it to your billing and compare it to the lp. I can see in various places how that data was captured and whether that amount makes sense or not. So whatever, whatever the AI is pulling out and decides whether there is like an issue here that Needs to be. I can check it with other sources of data in the enterprise. So because the data is structured then I can augment the results with rules and checkpoints to make sure that it resonates or it doesn't resonate.
- [18:48]
A
Have you crossed a million dollars of ARR at this point?
- [18:50]
B
Yeah, we crossed a million dollars of ARR. We are about at like 1.5 again. We, we just started to sell last year. Every engagement is kind of significant and yeah, we would like to here.
- [19:03]
A
What is your prediction? What would you like to end 2026 at in terms of ARR?
- [19:07]
B
I think 2026 is going to be the day yield for AI in the enterprise. I think we can triple our business
- [19:14]
A
get get up to 4.5 of ARR. Hey, as we wrap up here, I know you're short on time that you know, public markets literally today are getting crushed in terms of the SaaS companies. When you're building a tool like this, you know, if people don't know you well, I mean I know you at this point, but if they didn't know you're a successful entrepreneur, you've been in tech a long time, they might go, ah, he's just building another AI wrapper directly from your mouth. What is your deepest moat that would prevent Claud or any of these other tools from launching something that would cause your customers to churn.
- [19:39]
B
Right. So we actually this is like feedback. At the time we heard from one of the largest VC fund on the planet where they said that in order to allow AI to run successfully in the enterprise, you have to provide him with reliable data. We have created unique technology exactly like I explained, like creating the foundation of data. So we created a unique, we built this unique, unique graph technology allowing us to pull all the data from the various sources and normalize the data in a way that allows AI to operate across all the data. You can't just run AI on each and every system in the office of the cfo. You have to give it the context of the end to end data from all the systems that participate in the business process. Again, as I said, whether it's like the quoting system, the CRM, the billing, the erp, the banks, whatever. Only when you give the AI the full context and you arrange for him the full context he can provide. And that's why you build it with any other tool.
- [20:37]
A
What I'm hearing you say is you have real process ip, you know, the etl, process, extract, transform, load. When you then normalize the data set, only then can you start to Talk to it and get real value from it. However, we're seeing, you know, folks like DeepMind at Google put out things like AlphaFold, which has a, I mean, a very complex thing to sort of, you know, codify and decodify protein strands. Why couldn't they apply that very complex sort of structure to what you've done and, you know, replicate the ETL process?
- [21:02]
B
Yeah. I think uniquely in final, you have to be able to create an audit which is a virtual for I would say since accounting existed for 500 years, that was a virtual term meaning how can you show the start and the end of each transaction. Now, we created that. That capability requires you to develop unique technology or ETL capabilities that we created with the graph. So this is not about a system that can see multiple data sources in the. And you have to create that. You have to link the data and provide a level of accuracy to corporate finance, which has to deal with again, again, leakage issue, compliance issues, and also the visibility to the auditor that no
- [21:44]
A
other approach can provide, which was approved by John Smith. And sort of this, what you're saying is like this spider web is actually your ip, your ability to tell a story.
- [21:53]
B
Exactly. But more than that, in the, in the world of the office of the cfo, you have to be able to demonstrate that you can link the data. What I mean by that, you have to provide the customer the audit trail to make sure. Because let's say you have the most sophisticated models that looks across various data structures and give you an insight. You cannot predict if on the next anomaly you will be able to identify the same insight. Because we connect the data, we have the. We can, we can actually show the predictability or we can show the quality of the insights that we're going to provide because the data is linked. So you, you have to first physically link the data to make the customer. And also if auditor comfort that we see everything linked ICOM you got.
- [22:39]
A
You got 3 million bucks of extra rooming around. I want to invest.
- [22:42]
B
Let's talk after this call. You have my.
- [22:44]
A
I love it. It's tough, right? When you push on these things, you start to understand the rapper and who's got real ip. I mean, it's clear.
- [22:50]
B
I mean this is, it's real ip, it's unique technology. We are actually doing additional crazy things that I didn't talk about because we have to keep some of the motor under the hood for now, but it's really crazy. And I think for the first time we're solving a very complex problem in the office of the cfo, which is a comple complexity of the data core system and the in the natural gap between the ability of accountant to deal with data and the complexity of the data. Because accountants need to make contact decisions, not data decisions.
- [23:19]
A
I come on that note. If people want to follow your story here in 2026, 2027, where's the best place for them to find you? Online?
- [23:25]
B
I come at SafeBooks AI or LinkedIn. I come Kaufman.
- [23:28]
A
SafeBooks guys booking, billing and revenue all align. This guy knows what he's doing. You better pay attention. Sold his first company for 360, almost 400 million bucks to intuit. Now he's at it again. He's got first line of code built for SafeBooks in 2023. Raised a small 4 million round. 2024 kept growing. Raised a little more in 2025. Closed their $15 million seed round and launched their first paying customer. We love celebrating paying customers now. Today, 15 paying customers past 1.5 million bucks of ARR. We're recording in February of 2026. His goal over the next 11 months break 4.5 million bucks of ARR. He believes he can do it. He's got the cash. He's got the vision Graph technology again. Agentic revenue integrity. Big customers paying 300,000 bucks per year. Check it out at Sai Ahikam. Thanks for taking us to the top.
- [24:11]
B
Thank you so much for having me me today.
- [24:12]
A
You won't believe this. CEOs revenue. Click here to watch the next episode right now.