Loading summary
Transcript30 lines
- [00:01]
A
When you run AI for marketing or sales and let's say 80% is correct, then that's good enough for marketing or sales. In finance, it's not good enough.
- [00:14]
B
Welcome back to another episode of Builders. As always, this show is brought to you by Frontlines IO, Silicon Valley's leading B2B podcast production studio. If you're bringing technology to market and want to learn from your peers, we have a library of more than 1200 interviews with Venture backed founders and marketers. Where they talk, all things go to market. And of course, if you want to launch your own podcast, we offer podcasts as a service to more than 80 tech startups. The idea there is very simple. You show up and host and we do everything else. Now with all that said, let's jump in today's episode. Our guest today is Haikam Kaufman, CEO of SafeBooks. Haikam, welcome to the show.
- [00:51]
A
Thank you for having me. Looking forward to our conversation.
- [00:54]
B
Of course. As I was just telling you pre interview, high expectations here. Whenever we bring on Israeli founders, they always bring the goods, they always deliver. They're always a lot more open, I would say, than the folks in Silicon Valley. So very, very excited to have you on.
- [01:06]
A
Thank you. Your wish is my command. I hope I'll meet your expectations.
- [01:10]
B
Imagine you're on a flight from Tel Aviv to New York and someone next to you is talking up storm and they say, what do you do? How do you answer that question?
- [01:18]
A
We do finance operations automation. So basically today people are deployed in the office of the CFO and CEO to do complex data centric activities in order to validate financial data, govern financial data. And basically we got to a point where a machine can fully emulate every activity that a human being can do in finance, in corporate finance, across structured and unstructured data, just like human beings, so being able to read the data, make decisions, so on and so forth. Of course we need eventually a human in the loop to sometimes make final determinations, but the initial analysis could be similar to it. That's what we do.
- [02:04]
B
In your background, what made you say this is a problem that you want to go out and solve? Why this problem specifically?
- [02:10]
A
So I actually started in the office of a cfo. My background is in was in corporate finance. I worked in a few companies, including public companies. I had the chance to work in four public companies and some of the inspiration I got was first working for a company called Mercury Interactive, which invented software testing automation. And the idea would be that in order to automate software testing, you would basically emulate people's human work in checking various pieces of the software. And basically that not only allows you to save, test some people in the loop, it increases your quality level, reduces mistakes, accelerate the launch of new software and so on, so forth. And I think today it's undisputable. Another thing that inspired me is that after we sold Chev to a company called Intuit, at Intuit, we had to develop automation to be able to govern some of our financial data because we were moving money around. And that also inspired me to think, you know, if one company needs that, why not other companies need that? And I think the power of AI, especially even in the last 12 months, has got to a point where we can really and safely, that's why we're called SafeBooks. We can really and safely fully emulate A lot of the activities in the office of the cfo, which are from an automation standpoint, are very different from a lot of other automations and because of the level of data integrity, accuracy and completeness that you need in order to meet financial compliance and the requirements.
- [03:47]
B
If you just look at the office of the CFO today, how would you describe the sentiment around AI in general AI tools? What does that look like?
- [03:58]
A
Okay, so we've looked at several researches that have been done very recently. We actually have a slide on that in our deck. All of you Dec, where they talk about, I think 2026 would be the year that the office of the CFO would start to deploy AI in a massive scale. But I think most of the people, if you ask them, according to those researchers, let's say above 70%, are still concerned about the level of accuracy and completeness of the answers and the data. You can't afford to hallucinate, you can't, you know, when you run AI for marketing or sales and let's say 80% is correct, then that's good enough for marketing or sales. In finance, it's not good enough. And basically what SafeBooks has invented is data infrastructure that allows AI to run at scale with full accuracy and completeness across your data and across your system. So the way we do it is we actually look at every transaction end to end across the system to make sure we are not missing something in the process. And that's what unique about us.
- [05:03]
B
So, and obviously it's in the name, as you alluded to there with SafeBooks. But what do you do to build that trust? You know, I think as individual users, I'm sure that you know, myself and you and everyone in the office of the CFO you know, those experiences that we have with ChatGPT where it just, you know, create something out of nowhere or it hallucinates. What do you do to really build that trust with buyers? Because like you said, it's not marketing. Like, you can't afford to be off by even a little bit.
- [05:30]
A
That's a very good question, which I was never asked before. But I think, first of all, many customers would run their own UAT where they compare the results to what they kind of like doing today manually. So that's one thing. The other thing is we are maintaining a certification called SOC1, which means that we actually get certified for the accuracy of like, the agents and the controls, which is uniquely required for finance. And on top of that, I think from when we show you stuff on your data, and most people know their data really well, we actually get to a point where you see stuff. We basically, you always have an aha moment of things you haven't identified before. So even if you run on your historical data, which is always important, finance and show you stuff which you may have missed in your historical data, and that's like give you a sense that there's a machine that they can do. And again, we are not fully removing the human in the loop, but we are reducing that manual footprint which sometimes consumes time and is error prone and all of that.
- [06:38]
B
So this show is brought to you by Frontlines Media, a podcast production studio that helps B2B founders launch, manage and grow their own podcast. Now, if you're a founder, you may be thinking, I don't have time to host a podcast. I've got a company to build. Well, that's exactly what we built our service to do. You show up and host, and we handle literally everything else. To set up a call to discuss launching your own podcast, visit Frontlines I.O. podcast. Now back to today's episode. What about the market category? So I see online that you have the title of the world's first financial data governance platform. If you just think about the customers that you serve today, are they in market, are they looking for a financial data governance platform? Or is that something that you have to go out and create that demand and then capture that demand?
- [07:23]
A
Another excellent question. I'll tell you why we started with financial data governance. Actually, if you Google it, we should be the first unsponsored result. Because I think that at the end of the day, you know, if CFOs lose sleep at night, it's because of the accuracy and completeness of their data. It's important in so many ways. It's important for your business results, it's important for your auditors and your audits, it's important for compliance, it's important for your financial planning analysis. But today because of like we are fully agentic, then another category we may claim is Finance Operation Automation, which is basically a way to say that everyone is doing governance today, but they're doing it manually for the most part, right? They're pulling data from different sources, they're reconciling it, they're checking it, they document it and fake hook does all of that. But maybe a nicer way to frame it today is Finance Operation Automation or gentic Finance automation.
- [08:26]
B
And when you look at that competitive landscape, you don't have to talk about specific companies, but what are the buckets that you would put companies into? So let's imagine your icp, your ideal customer. What are the buckets of options that they have if they're not using you? Right?
- [08:42]
A
So I'd like to think that in the agentic finance operations you see maybe two or three type of companies today. You see companies that have like essentially a niche solution for transaction processing automation. So like AP, AR, billing, whatever. So each of those would have its own solution. Let's say Billing would also have work automation or AP would have like all expense management have their own automation for anomaly detection and things like that. What you are lacking today is a system that sees everything end to end. What I mean by that is if you think about your revenue processes, they involve your cpq, your quoting system, your CRM, your documentation, which is sometimes on Ironclad or on Salesforce or whatever, your billing system, your ERP forever and the other journal entry and your banks. So I'd like to think that SafeBooks is the first system that using our like I think AI powered graph technology, basically ingest all the data from all these systems, normalize it, arrange it in a way that the agents can see each and every transaction end to end. We literally create the very first audit trail which is like the foundation of disciplined finance. And by that we make sure not to make any mistakes. So today you have like, you know, AI powered like billing solution, AP and so on and so forth. That's kind of like one segment. Another segment is like close automation. So you know companies like Flowcast in America and Blackline and Ledge and others which basically manage the close process in, in the process of doing that, they offer some automation mostly around inside the ERP reconciliation, but looking at the entire process holistically, like human being, I think that's like a third new category that we are doing. And that technology was very complex to develop. It's a known problem because when you think about even like we mentioned, the order to cash, the revenue process or the AP processes, so you know you have many systems participating. In each of them the data is structured, but altogether it's all unstructured because in each of them it's structured differently. So they do talk to each other. But the way the quoting system is structured is different than your salesforce is different than your document, which is like an unstructured piece of data different from your billing system. And how do you make sense in all of that and arrange all of that in a way that a gentic workforce can do a good job is kind of like our moat.
- [11:24]
B
This show is brought to you by the Global Talent Company, a marketing leader's best friend. In these times of budget cuts and efficient growth, we help marketing leaders find, hire, vet and manage amazing marketing talent for 50 to 70% less than their US and European counterparts. Part to book a free consultation, visit globaltalent.co do you think that agentic is going to be a phrase that we're all using three years from now? I feel like that's very much in the vocabulary of tech right now. Is that one of those terms that's going to eventually fade away or do you think the word agentic is here to stay?
- [12:00]
A
You know, at Mercury we used to have a cmo, Chris Lockheed, who's actually one of the my investors and advisor.
- [12:07]
B
This is the category design guy.
- [12:09]
A
Right? Right. Yeah. So I've been working with Chris for 20 years and he used to say in China it's called food. He doesn't call it chagisu. So I'd like to think that it will go away and it will be just be like, you know, it will be so natural. Right. I think 30 years ago every software would be considered automation. Right. Because there was a sense of like something powered by a computer. So I'd like to think the agentic is like a way to say now that the software or the model can look at data, understand it, make a decision and act on that decision. But I think that's going to be so natural that, yeah, I think it's like one of my analogies is autonomous driving. Right. So I think a lot of maybe Tesla and others would say we have autonomous driving. I think in five or seven years every car would have autonomous driving.
- [13:00]
B
So I've read play bigger many times. I'm sure many of the founders in the Audience, you know, it's kind of like the bible for a lot of folks. Even just the other day, someone texted me saying, like, oh, my God, have you heard of this book? You know, it summarizes kind of everything people believe for you. You know, you're in a unique seat that you've actually worked with Chris and you known him for so long. You gave a good example there of in China, it's food. What else have you learned from working with Chris and just from knowing Chris and being around Chris?
- [13:24]
A
Yeah. So Chris, actually in Playdinger, he talks a lot about one of his case studies in Play Bigger is how Mark Benioff kind of like, started Salesforce. And I think while it's a great example and a great inspiration, he talks a lot about at the beginning of a company, when you try to define a category, talk about the problem, not necessarily the solution. And we've been trying for a long time now in our social channel to talk about the challenges today of, you know, you have, like, an exodus of accountants. People don't go, you know, in massive scale, right? You have, like, the increasing requirements from the regulatory authorities. You have the increasing complexity of systems and data. Data remains the number one issue. So keep talking about that. And when you talk about the problem, then we see people relate to that. And when they relate to that, we find them reaching out to us and wanting to kind of like, share their views or hear what we have to say. And that's one of the other things I've learned from Chris about that. Just talk about the page, talk about the problem, and then, you know, hopefully the solution will kind of evolve from there.
- [14:31]
B
And final question for you before we wrap up here. If we just think about your company, your product, your platform, what's the big picture, and what does the platform look like three years from now or five years from now?
- [14:42]
A
I'd like to think that most like we are seeing today in coding, most of the financial activities, the term close event would go away because close means, okay, I spent the last 30 days or the last quarter like, doing a regular job. Now, let me pause here. Look at everything, check everything. I think close would happen on a daily basis. I think many of the technical activities around finance operations will go away. The systems will remain. The systems of records will remain. Right, because you need a source of tools for. But the entire work around them would be fully automated and there would be human in the loop to. Obviously, like today, you have engineers who manage agents. You would have people who would manage energetic workforce that will take care of all these tasks, I think in a much faster pace and a much higher accuracy than people today. And I think accountants would then be able to focus on the, the core skill set, which is making accounting decisions, you know, writing the reports and all of that. But the whole specific, repetitive technical work around data around reporting would go away completely.
- [15:58]
B
Amazing. That's where we're going to end things. I would love to have you come back on in about a year or so just to keep us updated on the progress but of what you're building and love how you're building it. Before we wrap, for anyone listening and that wants to follow along, where should we send them to? Where should they go?
- [16:13]
A
Best is to go to our LinkedIn page or our website. So it's SafeBooks AI or the SafeBooks AI page on LinkedIn.
- [16:22]
B
Amazing. Well, thank you so much for taking the time. It's been a lot of fun.
- [16:25]
A
Thank you, Brett. Thank you for having me.
- [16:29]
B
Well, that's all for today's episode of Builders, brought to you by the Frontlines. If you want more amazing content like this, visit Frontlines IO, where you'll find a library of more than 1500 interviews with founders, marketers and other GTM leaders, where we unpack the tactical lessons from their journey. And of course, as always, if you do want to launch your own podcast, we'd love to have a conversation with you. Visit Frontlines IO Podcast as a service. Mention that you listen, mention you love the show, and we'll give you a 10% discount. Thanks for listening. We'll catch you on the next episode.