Kareem Khattab from Entendre Finance on AI and Blockchain in Accounting

[00:00:00] Umar: Welcome to The Accountant Quits, brought to you by the Web3 CFO Club, a community by Request Finance. With a curated community of Web3 CFOs from companies like Aave, The Sandbox, Binance, Consensys, Ledger, and many more, joining this club will allow you to network and learn best practices on web3 financial operations.

[00:00:24] Umar: On The Accountant Quits podcast, we discuss how blockchain will impact the accounting profession and how accountants should prepare themselves for the future of work. My name is Umar, your host, and even if some might refer to me as the accountant gone rogue, my job is to provide you with the blockchain knowledge that you need that will be relevant for the accounting industry as a whole.

[00:00:47] Umar: Welcome to Episode 59. Blockchain at its core allows the transfer of value without an intermediary and offers an immutable ledger that provides real time access to financial data, while AI based applications can be used to analyze financial data and automate repetitive tasks like data entry. A powerful combo for accounting.

[00:01:09] Umar: We are on the brink of a fundamental transformation and the impact on the accounting industry will be generational. So where do accountants go from here? Will AI replace accountants and 10 years from now, what will the accountants of today be doing? To help us find some answers through a discussion on the intersection of blockchain and AI in accounting.

[00:01:31] Umar: Today I have the pleasure to be speaking with Kareem Khattab, the founder and CEO of Entendre Finance, an AI powered crypto accounting software to help you automate your fiat plus digital assets activities. Last year, Entendre announced a $4 million seed round financing and since then has built a first of its kind crypto accounting tool with notable highlights such as self onboarding or chat GPT styled prompts to quickly analyze your on chain transactions. 

[00:02:01] Umar: Some of their users include Rain, Exponential Finance, Premier Finance, Roko Network and Hedgey Finance. In this episode today you will learn new opportunities with the intersection of blockchain and AI, implementing AI in accounting for a crypto accountant, an overview of Entendre Finance and its positioning against other crypto accounting solutions, emerging opportunities for accountants to upskill themselves with AI and blockchain and much more. 

[00:02:32] Umar: Kareem, welcome. And thanks for making the time to be here. 

[00:02:35] Kareem: So thank you for having me. I started my career as a software engineer at Workday building cash management, banking, settlements, and financial reporting products. And then I switched from software engineering to product management, leading payments and accounts receivable at WeWork at the peak valuation through the SoftBank Takeover.

[00:02:56] Kareem: Around that time is when I actually started getting really interested in blockchain, particularly because of our work around global expansion at the time, and trying to understand the intricacies of global payment stack and how difficult it was to build really lasting businesses in emerging markets.

[00:03:14] Kareem: And then I became a group product manager at CloudKitchens, which was a stealth company at the time I was working there as an early product hire and built out the entire financial platforms team there. And we actually started flirting with the idea of stable coins and how it can help our restaurant partners, which have traditionally low margins to undercut the card networks and, and put more money in the pockets of our restaurant partners.

[00:03:40] Kareem: And at that point, I was really, really excited about the opportunity of starting Entendre because of democratization of large language models. And because I knew that this was something that we struggled with at my previous high growth startups and knowing how fast crypto moves as an industry. 

[00:04:01] Kareem: I knew that this back office accounting and financial health management was going to be a challenge and at the time we were getting gearing up to to start Entendre. 

[00:04:12] Kareem: Some of the biggest, you know events within crypto happened around the UST crash, some of the more notable lenders like Celsius, Voyager, etc Three Arrows Capital, and obviously FTX.

[00:04:26] Kareem: So our, you know, thesis was pretty well timed that we were going to need more than just human capital to solve this problem, we were going to need to leverage a lot of the new advancements in artificial intelligence. 

[00:04:41] Umar: I want to start the conversation today with the topic of the intersection of blockchain and AI.

[00:04:46] Umar: Like I said in the intro, blockchain allows the transfer of value without an intermediary, and it offers an immutable ledger that provides real time access to financial data, while AI, on the other hand, it can be used to analyze that data and automate repetitive tasks. 

[00:05:04] Umar: Now, as someone building an AI powered crypto accounting software, what excites you about the intersection of these two technologies and how can AI and crypto help each other improve?

[00:05:16] Kareem: Yeah, absolutely. So traditional accounting software used to be on premise, very closed off and hard to obviously scale globally because you were limited by the geography of your, you know, database administrator and where you're storing this data.

[00:05:32] Kareem: And then obviously with the, you know, cloud based solutions coming to the forefront, people were still very much private about their financial data, but they understood the advantages of using the cloud and making that data accessible to the right people.

[00:05:48] Kareem: Now, obviously this idea that now everything is on a public ledger somewhere on a blockchain, you know, was scoffed at in the early days and now trillions of dollars of financial activity is happening on chain. 

[00:06:01] Kareem: And so, it's really interesting to see that evolution because it allows us to basically build more openly and to use technology that traditional accounting systems couldn't really leverage to make accountants lives easier. So, you know, one of the limitations that we had at Workday was not being able to analyze our customers data to provide insights and recommendations.

[00:06:24] Kareem: This was something that was built into the contractual agreements at the time. And that really inhibited innovation and product development and now that everyone's data is on chain, it is actually much more open and allows us to build more sophisticated experiences that honestly weren't couldn't exist in traditional accounting software like QuickBooks or NetSuite, etc.

[00:06:47] Kareem: So I think that intersection is what's the most exciting for me because it really is an unlock and I think it's going to supercharge companies for the foreseeable future.

[00:07:00] Umar: Perfect, I want to go through some of the different challenges of bookkeeping and reporting for a crypto accountant.

[00:07:05] Umar: The first is data collection from block explorer. One of the main challenges with aggregating data from these block explorer is ensuring the accuracy and consistency of the data. And if you're using a subledger, the accountant would have to run what I'm calling data sanity checks or perform reconciliation exercises.

[00:07:25] Umar: The first would be for the balance itself, and the second would be to perform a transaction count check. One of the areas of strongest impact for AI is around data collection. I want to ask you how can AI help to improve the data collection process for on chain transactions? 

[00:07:44] Kareem: Yeah, so these language models are extremely powerful.

[00:07:48] Kareem: And as far as data collection goes, I don't think everything has an AI use case. I think, you know, with Entendre's engineering, we don't need necessarily AI to help pull in data or normalize data. I do think there are very interesting use cases, you know, that you could apply to AI. But we're actually, we're not on the indexing level.

[00:08:09] Kareem: We're not, you know, trying to get in the blockchain indexing game, but what your point is about reconciliation, you can think of the language model as your analyst, right? And you can have it run the checks that you want just by describing your requirements. 

[00:08:27] Kareem: And I think avoiding large Excel exports and having to use pivot tables and constantly compare what's in your sub ledger to what's in your spreadsheet effectively can go away when you have a supercomputer or a language model essentially doing that kind of heavy lifting for you and pointing out, you know, the deltas in your balances or why you know you have imbalances on your financial statements. 

[00:08:56] Kareem: And it can recognize those things almost instantly. And I think after one or two times of validating it yourself, You can have a pretty high degree of confidence that it's going to give you consistent results. And I think that to me is probably the better use of AI when it pertains to accounting use cases versus kind of the piping, you know, the kind of the infrastructure to pull that data in.

[00:09:20] Kareem: It's incredibly difficult, obviously, to have all the chain coverage in the world on your own infrastructure. You're going to have to leverage other solutions for that. 

[00:09:30] Umar: So then the other notable feature of sub ledgers is around tracking transactions. So once you've imported these transactions into your sub ledger, you'd have to label them, create automation rules and have them mapped to your chart of accounts.

[00:09:42] Umar: Now, as a preface, I observe a lot of businesses, not necessarily in crypto, they are engaging in AI washing, meaning a lot businesses now they just wrongfully declare that they use AI solely to promote and sell their product or service. 

[00:09:57] Umar: In the context of tracking your transaction, can you explain how LLMs can improve the tracking of on chain transactions, automate the categorization, and let's say automate the creation of journal entries?

[00:10:11] Umar: And ultimately, the analysis part, I believe we understand it already. 

[00:10:16] Kareem: Yeah, of course. Yeah. So Entendre, at least when you think about use cases, accounting is almost the perfect use case for AI. And the reason it's even better for crypto is because of how fast crypto web3 moves. And so there hasn't really been a technology that can move as fast as crypto has besides AI.

[00:10:43] Kareem: And when I say AI is moving fast, I mean, it is almost like a daily thing that some new innovation or some new open source software or framework is being built. And that's incredibly exciting because every time you kind of jump in, you realize what, how much better we can make the customer experience be for, you know, our stakeholders.

[00:11:08] Kareem: And so I totally hear you on the AI washing piece. For Entendre, we're not just, you know, enabling an A. I. help center, you know, where you can ask it questions on documents or you know, only, you know, write, you know, write query data with natural language. Those are obviously extremely cool and helpful, but we're actually no longer building a sub ledger.

[00:11:31] Kareem: We're building an accountant and that didn't actually exist or wasn't possible unless you had a massive capital investments in training and building your own, your own language models. So for us, I think we kind of break it up into three parts. There's the onboarding of your data and getting that initial implementation complete.

[00:11:53] Kareem: It's the ongoing maintenance of the accounting system, your sub ledger and keeping it clean and not having to spend hours of your time making changes, configuring things. And then the third piece is the analysis. Well, now that Entendre is the system of record, we can start building you some really, really cool graphical analysis.

[00:12:13] Kareem: So we're really heavily invested right now on the accountants layer of our, of our stack. Entendre as a sub ledger, we've been pretty much code complete on that for quite some time. We actually haven't even live for more like we exited stealth last year. So we're about shy of a year of our official launch date.

[00:12:36] Kareem: All of our time and investment now is to make the experience even more delightful. We're, and that's really been the focus for us. 

[00:12:44] Umar: As a follow up question, there are a lot of crypto accounting solutions, but what is the main challenge today building a sub ledger? 

[00:12:52] Kareem: Yeah, I think interoperability with the cut that specific customers tech stack, financial tech stack and being modular. 

[00:13:02] Kareem: And I mean it in the sense that you have to build a solution that regardless of how the business is structured, right. Or what countries they operate in or what currencies they operate in, you can give them a uniform experience across all of their, their entities.

[00:13:17] Kareem: And I think we've done a good job with that, with Entendre, we've built those kind of underlying primitives needed for a sub ledger. Such that we're not really lacking anything when we are in the sales conversation with a customer. It's always fine tuning or tailoring something to their needs. That's adding a new chain or adding a custom integration, but we've never actually had to go back to the drawing board and figure out how do we fundamentally help this customer close the books. 

[00:13:51] Umar: Very interesting. Now, while preparing for this episode, I came across a term called AI agents. So, right now, interacting with AI, for a lot of people, that follows a familiar ritual. You enter a prompt, and then the AI model calculates a response based on the input, right?

[00:14:08] Umar: So, AI agents, as I read them, they work in a different way. They are designed to think and act independently. Can you explain to the listeners what is an AI agent and maybe provide an example in the context of crypto? 

[00:14:23] Kareem: Yeah, absolutely. So I like to think of AI agents as a script. You're giving a language model a defined set of instructions that it needs to do and will not stop until that task is complete.

[00:14:40] Kareem: And the concept of stringing along those processes will be super applicable to finance, accounting, and obviously subledgers. I think that's almost the future. You can almost say if Entendre right now is building the AI accountant, at some point we need to build the finance systems analyst, right? Or, you know, the FP&A analyst, right?

[00:15:05] Kareem: Like, those are the people who are going to, you know, be able to create these flows for the language model to follow so that they can prepare management reports faster. They can enable, you know, new integrations or respond to, you know, internal tickets very fast, right? Those are the type of things that I think also make managing and implementing accounting solutions, especially like large enterprise planning solutions extremely expensive. It's really the maintenance.

[00:15:39] Kareem: And so I think that is the opportunity in my mind of the power of AI agents. Now for crypto specifically, I think there's a ton of exciting opportunities. I mean, there's so much that happens on chain. You can almost create an agent for any particular string of use cases, right?

[00:15:59] Kareem: Like go and fetch me all payments made to this vendor with, you know, with this address and make sure that you, and these are like, you know, the net terms and make sure every time you receive an invoice that you, you send that payment out from your Gnosis Safe for approval. And then once it's approved, go ahead and fire that payment off and let me know when it's actually been complete.

[00:16:25] Kareem: Right. So there's so many things that you can string along just by explaining to this agent what to do. And I think Entendre, like at the current where we currently are as a, as a team, I think there's still so much left that needs to be built on the accounting front that we're not actually jumping into the operational side yet.

[00:16:47] Kareem: I think like there's still a ton of gaps to be able to support major, large businesses that were just solely focused on how can we make it so that Entendre is that one stop shop. And they can go ahead and pick their favorite tools or front ends to do all the other operational on chain duties. 

[00:17:09] Umar: So I'd like to speak a little bit more about Entendre.

[00:17:12] Umar: Entendre is a new player in the subledger market, but please do correct me if I should not refer to it as a subledger. And in light of what we've previously discussed about the intersection of blockchain and AI, can you tell us a little bit more about, or share an overview of Entendre finance and the features today that differentiate it?

[00:17:33] Kareem: Yeah, absolutely. So I think Entendre can, you can refer to Entendre as a sub ledger. We actually have customers that use us as their GL because they're 95 to 99 percent on chain. And so that's, I think that where that kind of like modularity comes in. Right. And I don't think we're going to see a world where fiat goes away entirely.

[00:17:59] Kareem: You know, if right now it's , a 5 to 95 percent split, and we know that crypto is going to grow the way that we're anticipating, you know, it's very possible in the next 10, 20 years that it's a 50, 50 split and that's okay. And so we want to work in a world where ultimately, at some point, we can support anyone's accounting use cases, but because of the difficulty it is to stay on top of all of the things that happen in crypto, it's not really worth your accounting team's time to have to be, you know, web3 experts right on top of the fact that they have to do everything else to run your business properly.

[00:18:38] Kareem: So for me, I think, you know, For us, we can act as a subledger or a general ledger, and it depends on your business. And I think I'm extremely proud to be able to, to offer that to folks from a subledger aspect, you know, we are multi entity, we support multi currency, and I think multi entity and multi currency were two things that we did early on because we knew how many DeFi projects have subsidiaries overseas.

[00:19:08] Kareem: And we know how, you know, DAOs versus labs operate and how they're structured, right? And, you know, how different projects have foundations and all these different things. So no one really, we didn't really see like a compelling solution in the market for how web3 businesses still operate like traditional businesses, right?

[00:19:31] Kareem: They want to be in the best tax advantageous structure. They want to be able to, you know, pass IFRS audits as well as, you know, GAAP audits and, you know, these are the type of things that we did early on and we didn't want to kind of build something that only, you know, fit one shape. We wanted to build something that can, can be pretty, can it be pretty expansive and that's something also that, you know, I learned in my experience working at these large real estate tech startups where.

[00:20:02] Kareem: Each property was essentially its own legal entity, had its own bank account, had its own P&L and balance sheet, and us having to kind of figure out the flow of funds and being able to extrapolate that out and build internal systems to support that, I think really inspired Entendre early design. So I think that would be kind of a little bit of an overview of entendre as far as how we approach the problem and happy to jump into more tactical features needed.

[00:20:30] Umar: Before we continue, we'll take a quick commercial break from our sponsor. Working in web3 can transform your career, be financially rewarding and surround you with a vibrant community. But as you're very much aware, this space requires rethinking a lot of the old models of how we work. For example, as the leader in a Web3 organization, it's up to you to figure out the most cost effective way to off ramp the company's crypto, or what's the most efficient setup to mass pay your contractors in crypto.

[00:21:01] Umar: Getting your organization to run on crypto is daunting if you're alone. That's why Request Finance, the industry leader in crypto invoicing, payroll and expenses, has curated a community of Web3 CFOs to share best practices around Web3 financial operations. With CFOs from leading projects like Aave, The Sandbox, Binance, ConsenSys and many more, joining this community will allow you to network and fast track getting your organization compliant in crypto. 

[00:21:32] Umar: And you know what? I'm also responsible for accepting new members and growing the Web3 CFO club. So if you're a Web3 business Founder, CEO, CFO, or in charge of financial operations, you can join this exclusive community today by filling up an application form at theaccountantquits.com/web3cfo. 

[00:21:54] Umar: Subject to a screening check, you will then start interacting with high profile web3cfos. get access to members only benefits like webinars, resources and invitations to physical meetups. Join the club today and let's win web3 financial operations together. 

[00:22:12] Umar: Yeah. One of the other questions I have is related to data privacy.

[00:22:16] Umar: How has Entendre AI feature been been built from that? I mean, Are you building that on top of Chat GPT and using their LLM? Because if you are. And I believe you'd be a customer of their API. Will the data provided by a client using Entendre be used to also train the LLM on OpenAI.

[00:22:37] Kareem: Yeah, so the other thing that differentiates us is our being the first mover in AI and really applying it to our system And this is actually really interesting. We're using a concept called function calling and what that means is we've built a very robust API which is very very important. When you think of some of the best fintechs in the world, Stripe, for example, Plaid, Ramp, I mean, their APIs are extremely well documented, easy to understand, but are very unique to their system and their infrastructure.

[00:23:13] Kareem: And so, it doesn't really matter how often you use one of their APIs without context on a person's business, like a Stripe customer's business and how they move funds. That data is just numbers right and to OpenAI or whatever LLM we , we're just calling our API using the LLM. We're not actually training a model yet.

[00:23:39] Kareem: We're not actually building something unique to our, you know, to accounting because accounting is a well defined set of rules and parameters, right? And it's actually up to the customer, if this should be a liability on their balance sheet, or it should actually go to expense, right? It should be like a P& L item or a balance sheet item.

[00:23:59] Kareem: That's super unique to each and every individual customer. So our goal here is to make sure that we're focused, hyper focused on the experience of making sure that the user can get what they need to get done. And we use that concept of function calling, which is just. using the LLM as that intermediary to make decisions and classify the user's intent extremely fast and then use what Entendre has to offer to accomplish it.

[00:24:25] Kareem: And so that's really the version one of what we're building. I think absolutely in the future, we're going to need to build something a little bit more robust. I wish I was an AI expert or an AI engineer. I am still learning, but we do use OpenAI today. But we are also flirting and opening up our infrastructure to open language, open source language models, because a lot of them are actually catching up to GPT and in some instances are actually faster.

[00:24:58] Kareem: And I've been geeking out on this new technology, uh, LPUs versus GPUs. And LPUs are proprietary to a new startup called Groq with a Q. So it's not the Twitter Groq, but Groq is fantastic. And their chip is actually meant to optimize the throughput of language models. Where GPUs are a little bit more, I mean, they're meant for graphics processing.

[00:25:23] Kareem: So there's obviously a lot more in a GPU than an LPU, which then can, you know, hurt that throughput that a language model needs. And so these are just things that we're looking into to, you know, be able to make the, the, the experience even faster than we thought possible. 

[00:25:41] Umar: You do have some  snapshots of, or some examples of using chat GPT on Entendre, but for the listeners, could you provide us some examples of what kind of prompts they could be inserting into Entendre?

[00:25:54] Kareem: Absolutely. Yeah, no, that's an amazing question. Some of my favorite are actually the configuration prompts. So you can quite literally drag and drop your chart of accounts and it will create a one to one mapping in Entendre for you, or you can even copy and paste say your wallet addresses and say, Hey, I want you to import these 20 wallet addresses.

[00:26:15] Kareem: And if you didn't specify the chain, it will ask you, right? And so it's even, we've even built it in a manner that like, if it's missing information, it won't make assumptions, it will ask you and you can have a conversation with it. Now on the analysis side, obviously, that's the most understandable side, but because Entendre is a system of record and actually has all of your double entry journals.

[00:26:37] Kareem: You don't even, you can ask it a combination of operational queries, like how much transactions do I have by chain, but you can also ask it really deep financial and accounting analysis, you know, like how's my income changed over the last six accounting periods. And it will give you the actual month over month change in a graph of your choosing 

[00:27:01] Umar: as an accountant, what I'm thinking while preparing for this episode, I was just trying to envision the future of it, where, of course, it can be used for analysis, but how much of it can be useful input for entry, for example, I wish I could write, can you create an invoice to that vendor and I'll like, and I want to receive my funds in this wallet address using this currency on this blockchain and make it recurring every month, And once I receive it, it should be accounted as a marketing expense.

[00:27:30] Umar: Everything just upon which then automates everything until, you know, until you have it in your PNL, that would be 

[00:27:37] Kareem: amazing. Yeah, I think it's just a matter of time. And I think what's funny is a lot of us thought where the current like state of AI is, wasn't going to happen for another few years. That's how scary.

[00:27:52] Kareem: It's like fast, it's moving. And so there's also a great deal of anxiety on our team internally, because we just have like this lightning in a bottle and there are so many things that you can build that you, you know, that you just mentioned. And so how can we continue to position ourselves as like the AI sub ledger thought leader in the market?

[00:28:16] Kareem: And that's, I think, you know, we're kind of competing against ourselves when it comes to the AI thing. And yeah, it's a very fun place to be at right now. And yeah, I think on the data entry side, like we've already started those elements of create my wallet, you know, import my transactions from this day to this day, you know, add these five ledger accounts, create this additional legal entity, right?

[00:28:42] Kareem: All of these little things we're slowly building on top of it. And then I think that's where now we need to start stringing together those processes. Which is also super exciting. And we have the first concept of that using onboarding. And so, you know, Entendre today has kind of, we're going after that implementation solutions, architects of sorts, where, how can we bring down the cost to implement or switch to near zero such that we can actually pass the cost savings onto the consumer, the customer.

[00:29:20] Kareem: Because if we don't have to hire 10 solutions engineers or, you know, customer success people, I mean, that's a significant efficiency that can be passed over, you know to you know to benefit our customers 

[00:29:33] Umar: Awesome. Can you tell us a little bit more about this? So you recently launched this onboarding concierge service. Can you walk us through the process for onboarding for a new client using Entendre?

[00:29:45] Umar: And how does the implementation time look like whether I would have a hundred or a million transactions across different blockchains and wallets? 

[00:29:53] Kareem: Yeah, absolutely. So we've certainly stress tested the best we could, but you can have tens of thousands of transactions. And what we'll do is. We collect all of your data up front, give us your entity structure, give us your wallets, add your exchanges, add any other integrations.

[00:30:10] Kareem: I think you'll have some questions on that, but we also have an application marketplace where people can add different web3 platforms that they use to continue to hydrate the transaction data. And then from there, our onboarding concierge will give you a default implementation set up based on all of the things that it knows about your business and automate all of the transactions that it knows it or it has identified.

[00:30:40] Kareem: And so a very simple example of this is internal wallet transfers, right? Well, it knows that, you know, Wallet A and Wallet B be belong to the same legal entity in your, you know, organizational chart and all the withdrawals and deposits that go between those two can automatically be, can automate entries for it automatically.

[00:31:01] Kareem: And that's super exciting because that in itself, by the time you get to into the product, you know, automated, you know, 20 percent of all of your transaction. 

[00:31:11] Kareem: On more complex example that it knows is bridges. So it knows all of the major bridge contracts. And it understands that when you are, you know, sending funds across different chains, and then it automates the entries for that, another complex one that it knows.

[00:31:29] Kareem: If you've added multiple subsidiaries within the onboarding flow, well, that internal wallet transfer entry is now an intercompany transfer entry. And so those are the type of things that now, before even talking to anyone on our team, you just have that out of the box without having to lift a finger.

[00:31:50] Kareem: And I think that is an extremely exciting experience for people.

[00:31:55] Umar:  It is. Next. I want to talk about the positioning of Entendre against other crypto accounting solutions. If there is, I always say, if there is one tool that every crypto accountant must be using, that's a sub ledger. And that's why I love interviewing founders of sub ledgers on this show.

[00:32:12] Umar: Our listeners will be familiar with Bitwave, Cryptoworth, TRES and Consola. So it is a very competitive market. How does Entendre differentiate itself from these competitors. 

[00:32:27] Kareem: Yeah. I mean, one, I think everyone building in this space is an extremely smart product team and organization, right? Because you're not, I think accounting is a slog, right?

[00:32:40] Kareem: It's an extremely tough consumer base, right? It's, you know, accountants are extremely skeptical and they are, you know, they're almost like the bloodstream of a business, right? They have to gather, you know, they have to travel across all the departments and understand everything that's happening and, you know, provide financial statements right from that activity.

[00:33:04] Kareem: So I think for us, you know, it is a competitive market and we knew that going in, you know, we're almost like you mentioned the new kid on the block, but I think what differentiates us is obviously the tactical aspect of the platforms, right? And especially in enterprise SaaS, you can go and, you know, all these comparison charts and talk about how great you are.

[00:33:25] Kareem: But I think from a tactical standpoint, we've built things that we saw when we started that we're missing in the market, right? 

[00:33:32] Kareem: And that is what was, you know, exciting for us when we first kind of explored this path, we accomplished what we wanted to accomplish there on the tactical sub ledger product development side extremely quickly, and we exited stealth as soon as it was ready and on the kind of foundational side, you know, being that first mover in artificial intelligence and leveraging that to help accountants, I think is, you know, we have, say, a six month head start per say, but we're not satisfied. And we're going to keep, you know, pressing on the gas there to maintain our advantage.

[00:34:13] Kareem: I also think, you the fact that we built something that we don't have to continue to always go back and reiterate on from an accounting focused side is what is allowing us to go all in on the AI aspect of things and we've talked to you know, some of the most senior people, you know, in finance within a lot of people's networks across large, not just web3 companies, but web2 companies and we our system can stand on its own two legs and we're extremely proud of that.

[00:34:47] Kareem: And I think. The last and final differentiator is the product experience. It's very hard to get talented product people to come in and work and do accounting technology. And I mean, someone who's worked at large accounting platforms in the past. Yeah, it's, it is a brutal place to try and change things on, on your customers because they're so used to doing something a specific way.

[00:35:15] Kareem: So. I think for us, we built an amazing product experience and we've hired really, really good product people on the engineering side on and on the design side. So I think that to me is, is also super important is not just making Entendre delightful from like a functional perspective, but making it delightful to actually look at and use on a day to day basis.

[00:35:38] Umar: I wanted to ask you this question earlier, but what does entendre actually mean? So in French Entendre means to hear. So I was thinking maybe Entendre is about hearing everything that happens, happens on chain. 

[00:35:53] Kareem: It's actually a little bit more, this is a great, this is actually the best question I would say.

[00:35:59] Kareem: Okay. So. I also believe Entendre means double meaning in French as well. And also the double entendre when, you know, people are wrapping and, you know, stringing together words, that idea of like double meaning to me resonated with a journal entry in that you have a double entry journal and what is an expense to say, you could be income for me.

[00:36:24] Kareem: And so for us, it was kind of a play on words, a double entendre, but like. That's what I think the name comes from or stems from was where we're automating double entry accounting for you. There's two sides to, you know, every entry, right. You know, and we want to have an entry for pretty much everything that happens on chain.

[00:36:44] Kareem: So that's kind of the, the backstory. 

[00:36:47] Umar: I want to speak about some of the integrations that you have with these web3 players, like Loop, Rain and Hedgey today. So you integrated first with Loop, a crypto subscription management platform. Our listeners will be familiar to to Loop. We had Eleni they founder on Episode 42. 

[00:37:09] Umar: You also integrate with Rain a crypto corporate card platform and we also had Charles on episode 46 and Hedgey, a token vesting platform who I haven't had yet, but probably soon in the future. Yeah, for sure. Can you walk us through some of these integrations and particularly the Rain integration, because I've been using Rain and the automation rules that you would have for their corporate card expenses. 

[00:37:36] Kareem: Yeah, absolutely. So what we kind of approach all of these integrations that are web3 related, like they're not data import related, like exchanges or, or pushing data to another system, like a QuickBooks.

[00:37:50] Kareem: We're just grabbing their metadata because they're web3 native companies, all of their stuff happens on chain. So we just need that additional context from them. So if you're, you know, a Loop customer, well, you have subscriptions, but that's not on Etherscan. So, you know, okay, well, this payment went to this subscription.

[00:38:12] Kareem: Let me, you know, reduce my accounts receivable and, increase my, wallet balance on my balance sheet. Right. And, you know, versus it being, you know, immediately recognized revenue or, or, some sort of a grant, you know, income or, or staking reward income. Right. And so that's the delta, like that's the difference.

[00:38:30] Kareem: And yeah. So even with Rain, right, we actually do pull in their fiat corporate card expenses and you can write automation around. Well, what card it's for or what merchant it belong to or what, you know, you know, let's say a category like of transaction. Is it is it fast food or is it, you know, public transportation and then on the on chain side.

[00:38:54] Kareem: Well, whenever you're on chain collateral gets debited, you'll, you know, have that expense being, you know, hitting your credit card statement expense line, right? So that's typically, you know, how we think about these integrations and there's been a massive surge of demand for, for us to, to integrate other folks.

[00:39:16] Kareem: We have probably five more integrations in the pipeline. All of them are unannounced. We will be announcing our next one around ETH denver time so stay tuned 

[00:39:28] Umar: Yeah, perfect. I want to move on to a slightly different topic and speak a bit about the future for accountants now with AI I often read that AI can enable accounting professionals to focus on more strategic initiatives, such as financial analysis and decision making, rather than doing manual bookkeeping tasks.

[00:39:49] Umar: However, if I'm the CFO and have a team of five accountants, I probably not need all of them to turn into a consultant, right? So. As someone building an AI powered accounting software, where would you recommend accountants today to upskill themselves? And if I were an accountant looking to learn web3, what skills do you think would be relevant for the next 10 years?

[00:40:13] Kareem: Yeah, this is, this is like the moral, like big topic in every industry, by the way, I a hundred percent agree that accountants should really be strategic partners in a business. They should not follow the we work model where, you know, there was like 3, 400 accountants, you know, doing data entry. I mean, that's not a world that even the accountants want to be in, right?

[00:40:36] Kareem: I mean, people want to do impactful work and people want to do, you know, they want to move the needle. And I think accountants are well poised to do that. I think, obviously, we have to figure out how that looks like in a world where there is AI incorporated in all these business processes. And I don't necessarily think that I don't think if you're an accountant today, I think your, your, your skill is to be that stamp of approval of what the AI does. 

[00:41:07] Kareem: And I think one skill that would be extremely beneficial for all accountants to learn, not just for crypto, but prompting and prompt engineering, I think. Being able to be an accountant prompt engineer is going to make you a 10x accountant, right? Just like there are 10x engineers and, you know, different roles like that.

[00:41:31] Kareem: You can do some serious damage in a great way, I would say, by understanding how to communicate clearly and effectively to language models. And not just throw your hands up in the air when the AI language model or the response isn't what you want. I think you should figure out prompting best practices, build prompting style guides for your team to follow, figure out how you can help shape the AI policy within your organization, right?

[00:42:01] Kareem: And like how data should be siphoned based on roles, right? And I think all of these things are going to be super important when the CFO is trying to understand the AI strategy of his organization or her organization. So I definitely think that's where I would go with, as far as upscaling myself, 

[00:42:22] Umar: Now for the last topic of the day, Kareem, I want to ask you whether AI can accelerate the institutional adoption of crypto. So one factor standing in the way of mass adoption of web3 today is the lack of user friendly interfaces and experiences using web3 can be technologically daunting and requires to get familiar to many new terminologies.

[00:42:46] Umar: What are you looking forward to in 2024 that you think could accelerate the institutional adoption of crypto? 

[00:42:52] Kareem: I think institutional adoption definitely depends on the financial tooling, right? And I think everyone wants to get their beak wet. Everyone wants to jump in. And that's, at least for us, like, at least in my experience.

[00:43:06] Kareem: It was extremely frustrating as a product manager to want to adopt a new tech, but it was finance stonewalling the innovation because rightfully so what's the plan if you flip this switch on, how many more accountants do I have to hire or is this going to blow up in my face at month end? Like what is the process?

[00:43:28] Kareem: And so I think for us, you're absolutely right. I think we have to make, you know, this is why Entendre is making the investments it's making right in AI and onboarding right. How can we make it a no brainer for any business to say, Hey, we have to move to blockchain, we have to be able to accept payments globally with, and not have to pay exuberant amount of fees.

[00:43:51] Kareem: How do we. earn yield on our assets that are just sitting there, right? Like how can we have strategic partnerships with all of these very innovative startups without any crypto strategy, right? So like, these are the things that if entendre is successful we'll, we'll see that adoption happen at a much more accelerated pace.

[00:44:12] Kareem: And I mean, it's happening right. With, with a lot of the new, like recent news, but I do think there's a lot of people pulling their hair out right now on the accounting side. 

[00:44:23] Umar: Kareem, it's been a fascinating discussion. I want to be respectful of your time as well. As closing thoughts has there been anything today that maybe we didn't touch on that you'd like to share as closing thoughts to our listeners?

[00:44:37] Kareem: No, I think this has been awesome. I think what you're doing is extremely important for the accounting industry and it was an awesome conversation. So, and the Entendre question was great, like where the name came from. So hopefully it was a sufficient answer. It was, 

[00:44:54] Umar: I have the last question that I like asking to my guests before we leave is do you have a favorite quote or maxim that you live by?

[00:45:02] Kareem: Yeah, I think for me it's enjoy failure. I, you know, I think failure and trying new things and getting that response loop back extremely fast is better than success. Cause I think ultimately that's what a bunch of little failures or large failures end up leading to a very big successful outcome. So I think that's how my team thinks about it as well.

[00:45:26] Umar: It's a very entrepreneurial answer. I think, and I completely agree. I think the only thing that matters is that you ship and you learn from it. So you can learn to better iterate for next time. Kareem, it's been a wonderful conversation. If people want to reach out to you or they want to learn more about Entendre, where should they go?

[00:45:48] Kareem: So obviously Entendre.finance is our website. Feel free to book time with us there. Follow me and on Twitter at my name, but instead of an a, it's just Kareem with a v. I was trying to be cool and you know, 10 years ago and Entendre Finance on Twitter or also on Warpcast, same handles. But I am available anyone, anytime.

[00:46:12] Kareem: So please feel free to reach out. 

[00:46:15] Umar: All right. Thanks a lot for your time again, 

[00:46:17] Kareem: Kareem. And we'll see you for having me. It was a great, great conversation. Cheers. 

[00:46:22] Umar: I would like to thank everyone for listening to this episode. You will find all the links of the episode show notes and transcript on the website of The Accountant Quits at theaccountantquits.com. 

[00:46:33] Umar: Please note that this content is for general information purposes only and is not a substitute for consultation with professional advisors.. If you do know anyone who could benefit from the episode and you care about them, please do share the episode with them. All the episodes are available on Spotify, Apple Podcasts and Google Podcasts.

[00:46:53] Umar: And by leaving us a review and rating, you will support the channel and all your fellow accountants. In order to be notified each time we release a new episode, do follow us on Instagram and LinkedIn. We hope to have you with us next time. Bye for now.