Latest / The Accounting Podcast / Interview and demo with Botkeeper CEO Enrico Palmerino on Facebook Live

Transcript

Interview and demo with Botkeeper CEO Enrico Palmerino on Facebook Live · The Accounting Podcast

• Blake Oliver: This bonus episode of The Cloud Accounting Podcast features an extended interview with Enrico Palmerino, of Botkeeper, and his technical director, Justin DoBosh. The interview was recorded via Facebook Live on Friday, March 1. It was organized by my co-host, David Leary, to address outstanding questions from our previous coverage of Botkeeper on this podcast, as well as an article I published on my blog, entitled, "It’s not a bot — the truth about Botkeeper, the Google-funded, AI-powered “bookkeeper replacement.”" In [00:00:30] advance of the interview, we sent Botkeeper a list of questions, and also asked them to make a demonstration of the artificial intelligence in action. At the beginning of the interview, there is a brief demonstration of a Botkeeper bot that codes transactions in the QuickBooks Online general ledger. If you would like to see, as well as hear the demo, follow the link in the show notes. Welcome to The Cloud Accounting Podcast for a special bonus episode. I'm Blake Oliver- David Leary: And I'm David Leary. Blake Oliver: We are [00:01:00] joined today by Enrico Palmerino. He's the CEO, and founder of Botkeeper, as well as Justin DoBosh, who is the technical director. Did I get that right, Justin? Justin DoBosh: Yep. Blake Oliver: All right. Welcome. Thank you for joining us, Enrico, and Justin. Enrico Palmerino: Thank you for having us. Justin DoBosh: Yeah, thanks. David Leary: Yeah, thanks, guys. Before we jump in, I just wanna thank Veem for sponsoring. A little Veem story. I had to pay Blake for some hosting costs, so I create my bill in and QuickBooks Online. By the time I opened up the Veem browser, the [00:01:30] bill was already there. It already synced it- or opened up Veem in my browser; the bill was already there, and already synced. Because I already had Blake's email address in my QuickBooks, I didn't have to do anything but hit a button that said, "Pay Blake." It was on a Monday, and then, on Thursday, Blake got the money in his bank account [cross talk] literally, I did nothing but open up the browser, and click the button. The only way it would be easier is if Veem had a button inside of QuickBooks Online. It's a good way to pay business to business, for sure. Blake Oliver: I have to say, from the perspective of somebody [00:02:00] getting paid, it was pretty darned easy to go in, and set up my account. I think all I needed was my email address, and address - that typical information; linked up to my bank account, and got the notification that David had paid me, and got the notifications as it proceeded through the stages, which was pretty cool, because normally, I don't get notified if I'm gonna get a payment. Veem was good about notifying me. David Leary: It notifies on both sides, right? Blake Oliver: Yeah. David Leary: I'm notified, like, "Oh, it's been taken out [00:02:30] of your account." "Now it's been moved to Blake's account," and the whole process in between. My understanding, though, because it's your first payment, it took up to three days, but then, it's a lot faster once ... If I pay you again the second time, it's just faster. Blake Oliver: Oh, that's great [cross talk] It's always nice, when we try something out that a sponsor ... We try out the sponsor's product, and it works. David Leary: Yeah. Blake Oliver: Good job, Veem. David Leary: Yeah, so, if you guys wanna try out Veem, as well, go to CloudAccountingPodcast.com/Veem, V-E-E-M, and you guys can try it out, as well. without [00:03:00] further ado, welcome, Enrico; welcome, Justin. Enrico Palmerino: Thank you. David Leary: I think we should start probably with a demo, and jump in. Enrico Palmerino: Yeah, I'd love to do it- David Leary: A demo of Botkeeper. Enrico Palmerino: Yes. I'll let Justin take it over and put it on his screen. It's funny, my wife was joking with me that ... She's like, "You're a quant major. You probably shouldn't be tweeting on Facebook. You're certainly not a PR specialist. Good [00:03:30] luck trying to do a demo of ML, and AI in 120 characters." [inaudible] opportunity; really appreciate you guys allowing us to do this. What Justin has just thrown up is an example ... Just so you know, too, we scrubbed out a bunch of any other PII here in this data. What Justin's showing is QuickBooks, and an uncategorized expense that has come into QuickBooks. Essentially, [00:04:00] this client, you can see down below, they had previously been using AWS as their hosting provider. Then, Justin can keep scrolling, and showing the recurring AWS transactions. Intuit would be ... If you create a rule that said, "Every time something purchased by Amazon, or AWS, is hosting," then, certainly, it would categorize it for you. Creating/maintaining/updating those rules can be a [00:04:30] hassle, and a headache, especially if products change descriptions, vendors start selling new things. One of the really cool things about Botkeeper is, in this case, this client switched to Google Cloud as their new hosting provider, so, that transaction landed in the uncategorized expense bucket. Think about this, especially from Botkeeper's stance, we process millions of transactions for our clients every month, and [00:05:00] there's a lot of these uncategorized transactions that are occurring. Even on the rule set, we actually, for the most part, turn off a lot of the rules, and then just leverage our AI to actually do the mapping, because we find it's more consistently accurate. What will happen next is essentially this transaction gets pulled into Botkeeper, where you can imagine on this list, if this was ... Justin's just refreshing the screen, showing the transaction again, getting [00:05:30] pulled in. You can imagine that normally, and that's if we're grabbing all the transactions across all the clients, it could be hundreds, or thousands of transactions in this bucket. It would take a decent amount of time for a human to go in, and do the categorization, or look up this vendor ... Because, you know, GCP isn't a super-descriptive definition. You might look at the memo ... It just takes time to click on those things, do a [00:06:00] Google search, et cetera. Behind the scenes, what will start to happen ... We don't click the button yet. I'll give a little background first. Behind the scenes, Botkeeper has multiple machine-learning algorithms running, as well as natural-language processing. What essentially happens first is the natural-language processing looks at all the text, and it starts cleaning it up, parsing it out, trying to make sense of it, and storing it. Another [00:06:30] algorithm runs that goes, and searches the web for any, and all information, data related to this particular transaction; anything it can find, and then tries to make sense of that, as well; store that in the database. Another [cross talk] David Leary: -kinda what you do as a human. You see some charge, and the merchant name's weird. You're like, "Did I charge that?" so you start googling it, and you figure out, "Oh, yeah, I remember going to that restaurant." It's kinda doing that process. Enrico Palmerino: Correct. The goal is ... That's why they call it artificial intelligence, but really, it should be thought of as augmented intelligence. It's performing [00:07:00] like an action that you would do, in a similar way that you would do it, and then, basically assisting you to affirm that, "Yes, this is what I would have guessed it to be.". Not to get too deep here, but another machine-learning model runs against a client to determine the complexity level of that business - what variety of vendors, and new transactions are processing, the quantity, et cetera. Then, another one runs to [00:07:30] actually try multiple machine-learning models against this client to basically say ... There are many different models out there that are random-forest models; there are linear-regression models; there's clustering algorithms, et cetera; there's combinations of the two. Then, that runs to basically try to figure out which of the models, given this client's complexity score, and all the other data that we have, which model is gonna most accurately, or most likely - the highest probability - guess correctly what this thing is? Once [00:08:00] it's figured that out, it runs the final model to do that categorization, to go ahead, and click the button. What's gonna happen in a very short period time, where all of those things that I just described ... You'll notice that in the Category button, as soon as this finishes running ... This would be running against ... Just picture, like on a daily basis, running against hundreds, or thousands of transactions. It accurately guesses and categorizes this thing that it's hosting. Then, [00:08:30] what it does ... If Justin goes ... Well, actually, it guesses its hosting. Our humans review it to basically say, "Does this make sense?" They can do a very quick review, because ... Just imagine, there's a ton of transactions here ... Quickly review. They can quickly edit, or update. If they start updating, it's pulling, from the Chart of Accounts, the different categorizations that this client might have. They would approve it, if it was correct. In this case, it's correct. Approve ... Then, what's gonna happen is that's gonna sync [00:09:00] all of these transactions across all of our clients into that client's QuickBooks. When the client logs back in, they're gonna see that these transactions have now been properly categorized and classified accordingly. This is just one step of a machine doing something. The next step would be ... Actually, say even a prelude to this would be if you had submitted to us an analog, like [00:09:30] some analog data ... Let's say it lived on a receipt or lived on an invoice. The first step of the equation would actually be that OCR runs against that document. Just to be clear, Botkeeper is built in various forms of technology. We built machine learning; we built AI; we've built a robotic process automation; we've built data integrations for pushing data in, and out of different systems, as you see here. We've built decision [00:10:00] trees, and workflows, and events to alert our own internal staff that they should look at something or spend more time on something. In that analog scenario, we're using OCR to grab data off; we're using machine learning, and some of these models to basically contextualize, et cetera, that data; figure out what it is. After we've figured out what it is, it's a matter of, "Okay, if this is an expense that's already been incurred, it's gonna get booked. If it's an [00:10:30] expense that hasn't been incurred, and, say, it's a bill that needs to be paid, and like your scenario with Veem, the data would get automatically pushed into Bill.com, in the appropriate ... Because Bill.com is one of the platforms that we leverage for bill pay, we can push into Bill.com all the appropriate data information. It would have done a check against the vendors that were already listed in the Bill.com, which are synced with QBO to say ... Even if it wasn't properly entered the first time, the same thing [00:11:00] that did that contextual enrichment, just to parse it out, would say, "Hey, most of this vendor's name is exactly the same, and the thing that's missing, or the thing that is here instead makes sense." Like, Inc., like period. If it didn't have that before, that's not a major change, and that makes sense that this is the same vendor. Then, either updates with the "Inc." component, and syncs across, or leaves the vendor as is, and drops the "Inc." Then, the client in that scenario ... Once [00:11:30] that's approved, once that's been completed, now there's a bill basically waiting for that client in Bill.com, waiting to be approved, and paid, and it's in the appropriate approval workflow, because the thing that got in the appropriate approval workflow is a rule set that we built with the client, basically giving us, or educating us on their business, and the different rules, and processes that they follow. Over such a dollar amount, I want it to go in this approval [00:12:00] workflow, and be approved by these two people, et cetera. That, I think, quickly describes a few different scenarios of Botkeeper automating work. We obviously have ... The machine doesn't do everything, it's not ... And it's not perfect, which is why we have interfaces for our team to go in, view, correct, assess, affirm that it's right. Let's say, kind of the way to think about AI, and automation - [00:12:30] automation is really, really good at automating simple, manual, tedious, mundane, high-volume tasks. When it gets to a task that's complex, that's probably gonna be done by a human. The nice thing is if that complex task, over time ... This is why how much of a client's work is automated by Botkeeper, versus assisted by human, it varies from client to client, depending on what tasks that client's using [00:13:00] for Botkeeper. Are they using us to do everything, including collection calls, and emails ... Collection calls are definitely happening by human. The goal of Botkeeper was to give you a one-stop shop to deliver the full swoop, or full scope of bookkeeping services that you might need, and to use as much machine, as much tech, along the way, to get it done as efficiently as possible; basically making our humans that we have be superhumans. Now, the beauty [00:13:30] is the longer you stay with Botkeeper, even if you do give us complex tasks, if, over time, we can identify a pattern on those complex tasks, and basically start to see that there is some semblance of a rule here. We can then build that rule set into Botkeeper to start performing what was previously a complicated task, because we just didn't know it, or see it often enough, and make that an automated task moving forward. That's basically it, in a nutshell [cross talk]. David Leary: It was interesting that [00:14:00] you used that ... Really, Botkeeper's a one-stop shop for a suite of possible services a bookkeeper would offer ... It's an interesting way to think about it. Enrico Palmerino: Yeah, it's not just software. If it was just software, we'd basically be giving people like, "Hey, here's some code; good luck; run it yourself." It's gonna do part of the bookkeeping function for you, and not the rest. Whereas, we didn't ... The market feedback we got was ... Our issue, among [00:14:30] accounting firms, was there's not enough good bookkeepers out there. There's a lack of supply in bookkeepers, and an influx in demand. They were having a really hard time meeting that. "We're seeing a lot of attrition on the good bookkeepers that we do hire ..." If I have one bookkeeper, and I only have 10 clients, and you giving me a tool that makes that bookkeeper, say, more efficient, [00:15:00] but doesn't actually ... It doesn't actually reduce costs. That bookkeeper just doesn't have to do as much work, and there might not be enough work for them to do to leverage that additional time. The solution we were filling was if you lose a bookkeeper, you can plug Botkeeper in. If you're going through a high-scale mode, you can plug Botkeeper in, and let us augment the need for additional bookkeepers, so your firm can focus on hiring more skilled accountants doing more advisory, consulting, critical [00:15:30] thinking. You could say, "I just want Botkeeper to do this part of the bookkeeping process, and then, I'm gonna have my people do these other parts." Maybe the complex tasks, you want to keep it in-house, or in your firm. It totally ranges. We have small startups that never hire a bookkeeper use Botkeeper directly; ideally, they sign on to Botkeeper through one of our partners, and our partner is basically providing the bookkeeping [00:16:00] services to them by white-labeling and leveraging our platform. We've had large corporations, or large accounting firms also ... They lose one, or two staff; that puts a lot of extra work on the staff remaining; those people get frustrated; maybe they leave. They bring us in as a stopgap, and then kinda incrementally use us, as the business grows. David Leary: You did cover a question somebody actually had: Can I get this tech, and run it [00:16:30] on-site, myself? Obviously, no, but really, what you've built here, if I'm really paying attention correctly, is you have internal in-house bookkeepers under your roof, right? Enrico Palmerino: Yep. David Leary: You are making them uber-efficient. Enrico Palmerino: Yes [cross talk] David Leary: -instead of hiring a bookkeeper that does ... I think, in Blake's blog post, instead of a bookkeeper that handles 10 clients a month, or 40 clients a month, you're pushing up to what numbers, I think, was in that ... ? Enrico Palmerino: It depends on the business. You could be doing 10 clients a month that are [00:17:00] huge companies [cross talk] David Leary: That makes sense, yeah. Enrico Palmerino: It's hard to say what that is. I can say, you know, Blake and I had this convo. He was telling me, with his business, that he had hired this one person who was a super rock-star accountant that was serving 50 or 60 clients, and he knew he could never replicate that. He wasn't gonna be able to easily find that again. All of our people average 50 or 60 clients, so we're achieving ... That said, we have a Fortune 5000 [00:17:30] companies that use us. They have augmented probably the need for many people. Take that 50 or 60, and maybe even multiply it farther, because there are some very large companies doing a lot of processing [cross talk] cyborg, to your point. We're part human, part machine. The combination of both means it's better than just a machine, and it's better than just a human. Blake Oliver: To get back to the demo aspect, I can see [00:18:00] how the software can dramatically speed up coding of transactions from the bank feed, and how that can be superior, in many, many respects. For a number of clients, that's great, because that's all they need is cash-basis bookkeeping. What about the situation, where you really need those source documents, and the bank feed is just not enough? Can you show us, now, how that would work? Like you described in our interview, taking a snapshot, [00:18:30] or scanning an invoice, and then, you guys extract the line-item detail, and post that properly into Bill.com, for instance. Can we see that? Enrico Palmerino: The challenge of seeing that is it pretty much ... Any of the other players, and the data extraction off of analog, it doesn't happen instantly. It takes some time to extract, process, run, confirm, validate. Justin can pull [00:19:00] up this ... He could run a test on the front end, but it could take two hours for it to end up, because it depends on our team's approval schedule, like, when do they look at that transaction to affirm that it's correct? We don't let ... I have this philosophy that less-than-accurate accounting is not accounting. What we didn't wanna do is push things that were, for the most part, accurate [00:19:30] into QuickBooks. We wanted to wait until we actually had a human do at least an eye, like a once-over validation that helps improve the accuracy a little bit more and get it as high as we can possibly get it. Depending on when that person runs that approval process, and cycle is going to depend on when that data gets pushed into your QuickBooks. We try to make sure that, along all the transactions we're processing, that pushing is happening at least daily, if [00:20:00] not more. Blake Oliver: What about situations that are more complex than line items? Even Jan mentioned job costing. How do you do job costing for your customers? Enrico Palmerino: Similar to the way that a human would do job costing is we'd need to have some sort of job code. If the job code is communicated, either on the document, or in ... Some construction companies that use us send us an Excel sheet with information that has [00:20:30] job codes on it, or projects that are going on, or they communicate to us, like maybe not using an actual code, but they're using physical addresses of the properties that are being worked on. As long as you give us data that we can map a transaction to, we can do it, but we can't ... We're gonna run into the same issues that a person's gonna run into, where, if we see a receipt, or an invoice, and we have no information about it, and there's a bunch of transactions on there, we're not gonna know where to put those transactions [00:21:00] any more than the best person would. Blake Oliver: The algorithm is not then assigning expenses to job codes. A human is referencing some sort of document, and then coding up the transactions by job? Enrico Palmerino: No, the algorithm is not automatically trying to create a job code, or automatically trying to come up with a job code, without one being provided; but, if you provide a job code, the mapping to that job, [00:21:30] and that expense happens by machine. Think of it as [cross talk] a human would say, "Okay, if I have a ... I'm given a job code. I know to associate the expenses with this job." If the job code's located on that item, or document in a place that we already confirmed with the client was gonna be consistent, then we know to take whatever that number, or code, do a query across the jobs that are in QuickBooks, identify that job [00:22:00] that it should be associated with, and book it there. Blake Oliver: I really wish we could see that right now. Another example is paper checks. Unfortunately, many clients still write paper checks in this country. If I write a check ... Let's say I wrote a check yesterday, dated February 28, and I mailed that check out to my vendor. It's not gonna clear until March. You'll pick it up on the bank [00:22:30] feed, but how do you know that it's supposed to be dated in February? Enrico Palmerino: Back to the machine is not magic. It can only ... It can enhance what a person can do, if the human had the appropriate data in front of them, but if you cut a check, and you don't tell us you cut the check - you don't send us an image of the check, you don't send us an email telling us you cut the check - we're not gonna know to recognize that expense until it hits. Then, if it hits ... Even [00:23:00] then, we wouldn't know to even query you as to whether or not this should have been booked, or accrued in a month prior, the same way that a human bookkeeper wouldn't know to do that either if they saw an expense hit, unless there was some communication with the client that said, "Hey, every time a check clears, can you ping me, and I'll let you know where I want to recognize that in our books?" Say that didn't clear, or say you told us ... Let's play the other scenario out, which would be say [00:23:30] you told us you were cutting a check, and this is the amount; we knew the check was being cut. We'd have that booked in the month that you wanted that check, or that expense to be recognized. Then, in six months, say in six months from now, we'd set a rule set that basically said, "We don't see another transaction that matches this dollar amount within this period of time, follow up with client, and say, 'Do you want us to write this check off? Do [00:24:00] you want us to go another six months, and potentially recognize it, if it eventually gets cashed?'" A lot of this comes back down to ... This is where, like, the automation in that example is not even the machine doing anything fancy around how it's booking or recognizing it. It's the machine staying on top of the human to tell the human when to follow up, or check in on something, or identify that, "Hey, this looks like that check; double-check [00:24:30] whether or not that you know the dollar amounts match up. Is this the same check we're talking about?" Blake Oliver: If we're not gonna see anything else, we can probably stop the screen share, right, David? Okay, great. Veronica said something interesting that I'd like to bring up here. She said, "Bookkeeping is not working the bank feed. That's just data categorization. What else does Botkeeper do that is true bookkeeping? Bank reconciliations?" [00:25:00] Enrico Palmerino: The question leads to an even bigger question, like what is bookkeeping, right? There are plenty of companies out there that grab data off of a receipt and categorize that for you that call themselves automated bookkeeping. There are cash-basis bookkeepers that would call themselves bookkeeping. Are we talking about bookkeeping, or, are we talking about private [00:25:30] accounting? I think this is more of a matter of terminology. Where do you draw the lines? What is bookkeeping? Botkeeper automates various aspects of bookkeeping; our humans fill in the blanks. The end result is your books reconcile; you get the financial statements you want; you see the dashboards, and analytics within our platform that tell you how your business is doing, and whether it's trending, or growing, or needs assistance. The source documents [00:26:00] all live in our Botkeeper system, so you have that for if you go through an audit. You have an easy way to upload and get information to us. You can connect different data sources to Botkeeper. We allow you to integrate third-party apps to each other. Blake Oliver: It's clear to me that, based on your user reviews - and you do have many passionate users, and accountants who use Botkeeper, and are happy with it - that the end product seems [00:26:30] to be, in many cases, excellent. Enrico Palmerino: Thank you. Blake Oliver: It's obvious, right? Enrico Palmerino: Thank you. Blake Oliver: Good for you guys for doing that. There are plenty of terrible bookkeeping firms out there, and terrible accounting firms [cross talk] Enrico Palmerino: -only get better. Blake Oliver: My question is really about how the sausage is made. I think that that's what we really want to know, as bookkeepers, and accountants, is how [00:27:00] much of this is human versus machine, really, clearly, truly ... In my opinion- Enrico Palmerino: It's not an easy answer. Botkeeper doesn't make one sausage and sell a million of them. We make variety of sausages, because businesses aren't ... No two businesses are exactly alike. It's hard to say that ... If you looked at one business, and what we're doing for one business, and the [00:27:30] level of automation ... We have a free bookkeeping product. There are businesses that can use Botkeeper entirely for free [cross talk] can do that, and the economics make sense, and that's forever. That's not a short-term license. The economics only makes sense to do that, if you can fully automate what's being done in that case [cross talk] Blake Oliver: Let's talk about definitions ... Let's say I go back into ... I have a client still, so I do bookkeeping for that client, in [00:28:00] addition to accounting, and if ... Let's say all I did was code transactions in the GL, and that was the only thing I did for them, and I never reconciled the accounts, and I never sent them any reports, could I really call myself a bookkeeper? Enrico Palmerino: You'd be surprised. There are definitely people out there that do. Fortunately, that's why Botkeeper doesn't just call ourselves a categorizer of transactions. That's why we deliver the full suite of bookkeeping services. You could be a business that says, "Hey, that's [00:28:30] all I want you to do, and I'll do the rest." You could be an accounting firm that has a team, that you say, "Hey, I wanna do the invoicing." You don't have to use us to send invoices to your clients. You can do the invoicing yourself. You don't have to use us to pay your bills. You can [cross talk] Blake Oliver: You brought up the invoicing; that's a question somebody asked: who's doing the invoicing? Is it the bots? Is it the algorithm? Enrico Palmerino: If we're extracting data off of a sales contract that's predominantly unified, or standardized in some way, we [00:29:00] can grab that data, and turn that into an invoice, but we usually don't send that invoice out to the client. We just put it into QuickBooks, and it's up to you to actually do the sending, because revenue for a business is probably one of the most sensitive aspects of that business, and so they usually wanna put a human eye on it. I'd say we have ... Maybe it's like 30 percent of our clients use us to do invoicing, and the rest of them do the invoicing themselves, especially as the complexities of invoicing continue to grow. [00:29:30] Blake Oliver: Let's try to lay it out for the listeners, and the viewers, what exactly the bots are doing. By bots, I don't mean bots in quotation mark, where it's people, and we're calling them bots. I'm saying what is the bot actually doing? It looks to me like it's a more sophisticated coding algorithm that's leveraging ... It's pulling in from OCR technology. It's figuring out what kind of document it is, and it's putting that into [00:30:00] the subsidiary ledger, or the GL, wherever it's- whatever system it's supposed to go into. After that, that seems to be it. Am I wrong on that? Enrico Palmerino: What's your definition of a bot? AI [cross talk] Blake Oliver: Anything that is fully automated that doesn't require human intervention. Enrico Palmerino: Moving data from one system to another would be a bot [cross talk] Blake Oliver: What do I mean by a loose definition, right? Enrico Palmerino: -the bot, because a bot ... Most [00:30:30] bots that are out there, even like robotic process automation, and large firms, the kick-off of running the bot has human intervention. This is one of the challenges, that it's an understanding of actually the underlying tech. It's an understanding of the term, or definition of bot, or of AI, or of machine learning, or of natural language processing, or of robotic process automation, or of data integration. There are various technologies at play. [00:31:00] Botkeeper has built, and leverages many of these different tools in a variety of fashions, and workflows, and then stitches them together to get as much of the job done for a given client as possible. Then, yes, humans fill in the gap. We're very proud of our humans. We have skilled accountants; we have skilled data-validation people. We're a global company with a team in three countries, 22 states, and three offices here in the US. I think we've got a lot of diversity that we're happy about, there. Blake Oliver: Speaking of the [00:31:30] global nature of your business, that brings me to my second concern, or question, big question, about Botkeeper, which is The Philippines office [cross talk] Enrico Palmerino: -office also in Nigeria, too. Blake Oliver: You have an office in Nigeria? Okay, well, that one I couldn't find anything about on Google. Do I have to search Botkeeper Nigeria to find it? Enrico Palmerino: No, we actually [cross talk] Justin DoBosh: That's our engineer team; that's it. Only engineer happens there, and it's only a part of our engineering team. There's [00:32:00] no actual bookkeeping, or accounting going on there. Blake Oliver: Got it. That's different, and because [cross talk] because my main concern is about client data. I live in California, where the California Board of Accountancy requires that CPAs get written authorization and disclose any even potential offshoring of their clients' confidential information. Enrico Palmerino: You'll be very happy to hear that [00:32:30] all data sent to Botkeeper rests and stays entirely in the US. That's why, if you look at our contracts, we talk about all the data processing being here in the US, plus, we have a team remotely. They're accessing data that resides, and is stored here, from a remote destination, through our application, and through our hardware, and tech, that doesn't allow them to actually take the data over there. Blake Oliver: I understand that [cross talk] I understand that the data is in AWS servers [00:33:00] that are in the United States, but are you really telling me that when the data is then displayed on a computer screen, in an office in The Philippines, that that data is not in The Philippines when that happens? Enrico Palmerino: The data has not been downloaded and has not been sent to The Philippines. It can be-. Blake Oliver: But just by being displayed on a computer screen, there's a human being there in The Philippines, or wherever they are in the world, that is viewing that information [cross talk] David Leary: -by [00:33:30] definition, if anything goes from a server to a web browser, a copy of that web page has just been made on that computer. Now, you may go to the next page, you may clear your cache, and it goes away, but that's just the way the internet works, right? It's a copy of a file being put on somebody else's computer and shown on their screen. Enrico Palmerino: There is no data that actually will reside on a desktop. There's no way to take screenshots in The Philippines. There's no way to download, or store. There's no [00:34:00] phones allowed in our office building. There's biometrics that are required to get in. It is a paperless office; USB drives are disabled. There is intensive firewalls, and securities, and auto-logger ... Auto un-loggers, so it basically un-authenticates the individual from our systems to repeatedly have them log in and authenticate with new passwords. Data is continuously ... As it's being viewed, it's being viewed within our applications that are a closed network [00:34:30] that's within the Botkeeper framework, where outside sites are not being able to be accessed. Yes, they're physically viewing the data there, but the bits on that data ... No file is being actually transferred or sent over there that could be extracted off of a computer and walked away with. Blake Oliver: This is great to hear, and I'm really glad to hear that you're taking these security precautions, but-. Enrico Palmerino: Security is our number-one priority. Blake Oliver: Right, but why isn't that anywhere on your website? That's my question is [00:35:00] that nowhere [cross talk] Nowhere on the Botkeeper website could I find a single mention of offshoring of data, or labor. It doesn't matter if the computer code, or the bits, and the bytes reside in the United States if [cross talk] Enrico Palmerino: -that's the big issue with offshoring is that, if you look at a lot of these ... No names, but there is a company that was called out for basically sending a lot of data through Mechanical Turk, and that data was physically going [00:35:30] to India, where people had that data, and had access to it, to do with whatever they wanted. They could walk out of that building with data- Blake Oliver: Right. I understand you're taking these precautions, but if somebody is ... Information can be transferred in more than just bits, and bytes. It can be simply transferred by somebody looking at it on a computer screen. The information that you were telling me is being transmitted to me through this Zoom session into my brain, through my eyeballs, and my ears-. Enrico Palmerino: Just because [00:36:00] I can see you on my screen, doesn't mean you're in my office. Blake Oliver: Right, I know, but the data that you were giving me is coming to me, in Los Angeles, now, right? It's not staying with you in New York or Boston [cross talk] Enrico Palmerino: You can have this data, because you're recording it, and you're downloading it, but [cross talk] Blake Oliver: Here's the thing. The accountants that you have in The Philippines, let's say one of them has a very good memory, or one of them has a mnemonic memory, or a perfect memory. They now have confidential client information in their brain [00:36:30] that I, as a CPA, who signed up with you, never thought that a human being was gonna see, because, in my perception, based on your website, is that the data is all being processed by an algorithm. Enrico Palmerino: What confidential client information is seen in that screen you saw Justin showing ... Remember, we have US accounting teams, so the complex accounting tasks, those reside here in the US. Our US team, here, is the [00:37:00] one that are performing the complex tasks. We have a data validation team in The Philippines that looks at all the data that's being processed, and confirms that the algorithm is doing its job and getting it right, but it's only seeing limited amounts of data- Blake Oliver: What exactly are they seeing? I'm curious about that, because my understanding [cross talk] is that a human being has to do the bank reconciliation, for example, that your ... The bots don't do that, so- Enrico Palmerino: Yep, and [inaudible] here in the US pushes the button, and performs the reconciliation- Blake Oliver: Does The Philippines [00:37:30] team have access into QuickBooks Online files? Can they log in as a user, and manipulate data? Enrico Palmerino: They can log in through our application. Blake Oliver: Right. Can they see ... You open your sandbox QBO file, right? Can they go into my client's QBO file? Enrico Palmerino: I'll have to double-check with [cross talk] [00:38:00] Blake Oliver: -because it can't be that the only thing, they're doing is just ... If they're in your application, the one that processes the bank feed, and then they click Okay, what if they have a question? What if they are not sure if something should be coded right? They'd have to go into the GL, and look at the data, the history, to make a [cross talk] Enrico Palmerino: -or they ping our US team, and they ask our US team to look into it, and double-check it ... One of the things that they can do is they can say, "Hey, can you run this down? Can you check in with the client?" Blake Oliver: What exactly is being done in The Philippines, then, because I guess [00:38:30] I'm not clear on that? Enrico Palmerino: Data validation- Blake Oliver: Right, but what sort of ... Only data validation [cross talk] and in what applications? Enrico Palmerino: There are plenty of tools, whether it's your general ledger, whether it's an expense-report tracker, whether it's an OCR tool. All of these tools leverage offshore teams to validate and confirm the accuracy of the data that's being submitted, and to train [00:39:00] up, or train down that machine. We have a data-validation team in The Philippines, which our investors came out, audited, along when they were auditing our tech, to see, "Okay, what's this operation look like? How well have you guys securitized it? How well are you ...?" Blake Oliver: Sorry, sorry ... I just wanna make sure I'm clear. Like a bank statement, right? In order to reconcile the books, you need to look at the bank statement, and you need to compare that to QuickBooks [cross talk] What other ... Does the team in The Philippines have access [00:39:30] to source documents? They have to, right, to do the data validation? Enrico Palmerino: No. Remember, you watch the algorithm do its thing, and contextually look, and scrape across the internet to grab the information that it needed, in order to categorization, and classification. Then, they're looking at it to affirm that this makes sense. Blake Oliver: Let's say a bill comes in, an invoice comes in, or a sales order comes in. Is somebody in The Philippines looking at that? Let's say they're [00:40:00] validating the data. They're gonna look at what you guys pulled from it, and how your algorithm decided to code it. Then, they're gonna compare it to the actual sales order, right, to [cross talk] the invoice? Enrico Palmerino: -in limited view, so that's through our application, and our US team, if it's ... This is actually the point, you can pretty much assume that if it's a simple task, which the simple tasks are usually dealing with a limited set of data, or a data set that is more, call it ambiguous, or is not containing PII ... That data [00:40:30] is then being affirmed, or confirmed by the team in The Philippines [cross talk] Blake Oliver: Heather had a great question. Heather Smith said, "How can you do data validation without knowing the Chart of Accounts, and past entries? How can you possibly decide whether or not it's coded correctly, if you don't have access to the client file?". Enrico Palmerino: Because what we're looking for is anomalies. We're looking for things [cross talk] totally weird, or out of place. Feed Google, and for whatever [00:41:00] reason, our system said lease, a real estate expense. You'd say, "No, that doesn't make any sense. This can't be right. Chase this down; figure out what happened." Now that goes back off to a human. It's one of the reasons why we have a team of 20-something human accountants in the US, who are assisting with these more complicated chasing-down tasks. They deal with the, "Hey, the machine didn't get it right. Figure out why and follow up with a client to affirm or deny [00:41:30] the accuracy of the system.". Blake Oliver: I'm sorry, maybe I'm just an idiot, but I'm just not clear on what they have access to. If people in The Philippines have access into QuickBooks Online files of clients, and I'm a CPA, using your service, I want to know that. Enrico Palmerino: Okay. Blake Oliver: Right? Do you think I have a right to know that? Enrico Palmerino: I think if you asked, or had a question ... Said, "Hey, what is the team in The Philippines doing? Is our data going over to The Philippines?" These [00:42:00] are all questions that we are more than happy to answer, or walk you through ... We're happy to walk you through, too - if you're asking the question as an interested party, or an interested client - walk you through all the security protocols; walk you through how this data, and systems, and information works. The downside is what we can't do, and one of the things you were bringing up before was ... If we publish on our site all the security protocols, everything, how we do what we do, it's basically like publishing how we defend the castle. Now we're making it much easier for the people who wanna get in that [00:42:30] castle to try to figure out the ways that they're gonna get at it- Blake Oliver: Well, you don't have to give away all your secrets, but plenty of companies go through security audits. Enrico Palmerino: They do, but they ... Also, plenty of companies don't [cross talk] Blake Oliver: Have you gone through a security audit in your Philippines office? Enrico Palmerino: We're going through a SOC audit this year. Blake Oliver: Okay, so you have not yet. Enrico Palmerino: Yep. Blake Oliver: I'm just kind of flabbergasted that a number of CPAs have reached out to me and said that they didn't know that you had an offshore operation. How [00:43:00] can that be ethical, or how do you explain that? Enrico Palmerino: I think it really just boils down to Botkeeper is a global company. We have clients globally. We have employees globally. We have hired ... We're very proud of our diverse workforce, because you can hire really, really smart people outside the US. You don't have to reside here in the US. Because we're a tech-first company, we don't ... We're not a public accountancy; we [00:43:30] don't do taxes; we're not regulated by the federal government to have to publish the location of every one of our employees. If you think about it, when you buy software that uses a human element to affirm AI, because AI is an assisted process, right now ... There's very little AI, I would say, out there that's fully, fully autonomous. You just think of the driving car, right? If you buy a Microsoft license for Microsoft 360, there's probably a level of AI [00:44:00] to that license. It's stored in the cloud; it's coming down to you. That Microsoft sales rep doesn't say, "Before you purchase this license, let me talk about all the countries in which we have employees, and where they reside." You buy- Blake Oliver: Right, but because those employees aren't [cross talk] supposed to be looking at my client data, but they are- Enrico Palmerino: David Leary represents a company that does this exact thing. They're looking at your data, but you ... They're looking at your data to affirm that [00:44:30] data accuracy, and integrity, but they're not ... So long as the data is not actually being physically present in that country, or being able to be downloaded, you're having the same security, and the same governance of law that you have here in the US. We're governed under the same US legal regulations around security of data that any US company would be. Blake Oliver: I'm sorry, I just ... If it's not a big deal, then why don't you disclose [00:45:00] this on your website? Enrico Palmerino: We didn't realize it was a big deal ... Like I said, many tech companies ... AI is a new thing that's emerging into the accounting practice. I think, if you looked at industries, in general, most tech companies that have global teams, they don't paste the location of those teams on the front page of their site. There's actually companies that are fully endorsed by the AICPA. I think Ernest pointed [00:45:30] this out; one in particular is endorsed by the AICPA, and they have offshore teams doing validation. They actually even send data offshore. I'm not saying that we are against disclosing this info, we just didn't realize the accounting industry was so concerned about that. We have marketing actually already working on launching an updated version of our site that's gonna disclose where we have offices, where we have people. We'll even [00:46:00] include, I think, probably in our sales process, now, a component of that, that we'll talk through for their data security. What we really need is the help of the accounting industry to help educate us on how can we support and benefit you better. We need the industry's experts to say, "Hey, you guys aren't doing this," like you have. We want you to do this, and that's great [cross talk] and learn from it, and we're gonna try to become better. David Leary: In your [00:46:30] communications, it's obvious there's been a shift, a little bit. You're communicating this out, and you've gotten better about it. If I'm understanding your history correctly, you were working in the outsourced-bookkeeping space before you started Botkeeper; before it was a tech company. When you were working in that space, did disclosures like this ever come up? Was there any awareness of this before, or is this like until it started popping up, in the last three weeks, four weeks, it was just never a topic? [00:47:00] Enrico Palmerino: No ... For instance, we don't hide our Philippine office. There's a Facebook page about them. Think about it, if we were really trying to mask, or disguise the fact that we have a global team, we wouldn't be posting on Facebook photos of being in our global team. We wouldn't be doing press releases about the launch or opening of an office there. We wouldn't- Blake Oliver: Wait, wait, wait ... Let's be clear, here. The only thing that I was ever able to find about The Philippines operation was a Facebook [00:47:30] page that is ... The title is Botkeeper PH. It's not even written out, and a single tweet from January of 2018, saying that you opened The Philippines office. That was it [cross talk] Enrico Palmerino: Google Botkeeper ... Because you're googling Botkeeper maybe Philippine ... I don't know what Google ... I honestly don't know what Google results you googled, but if you google "Botkeeper Balanga," or "Botkeeper Invest in Balanga," there are news articles about us opening an office there. We're proud of it. We love our team in The Philippines; they're [00:48:00] brilliant people. They're hardworking [cross talk] David Leary: -group party. You had an awards dinner for them ... Enrico Palmerino: Yeah, we had a ... Like I said, we're not trying to hide this. You wouldn't publish, or post those things, and we were publishing, and posting those well before this all came up. If we were really trying to hide it, we wouldn't put it online. We wouldn't have a Facebook page for The Philippines. We wouldn't allow ... We wouldn't have Philippine employees on LinkedIn being associated with Botkeeper. We never tried to hide it [cross talk] realize [00:48:30] the accounting industry cared so much about the location of our employees. Blake Oliver: I would say there's a difference between trying to hide something, and not disclosing something, and being up front about it. I would like to call out a comment from Megan, who said, "You didn't realize people would care that the bots are actually people? I care." Enrico Palmerino: This goes back to what are you calling bots? The bots are [cross talk] Blake Oliver: It seems like your definitions ... Your definitions are very fluid - the [00:49:00] definition of what's a bot versus a human. What's bookkeeping, even. What's the definition of bookkeeping? How do you define it? Enrico Palmerino: I don't know ... David Leary, you represent a business, where, on their site, they say, "automated bookkeeping," and they extract data off of a receipt, and they categorize, and classify that. Now, is that ... David, you're an expert-. David Leary: I am gonna be really a poor example on this, and the reason why [00:49:30] is I've always ... My career, I've always used the word bookkeeper, CPA, accountant kind of all interchangeably. I know that pisses off CPAs, because they're like, "No, that's a bookkeeper ..." People can get very defensive about this. I think, in general, a lot of times, people think of bookkeeping as kind of the logging, and then maybe the reconciliation. You're definitely not doing tax, but [cross talk] getting it logged, or getting data shoved into the accounting [00:50:00] system, however that is - if it's manually typed, if it's done through some automation, or sent through a photograph. Getting that data in, and then probably a reconciliation, but then, maybe not much past that. Obviously, people are talking about, "Oh, you gotta be an advisor," and all this other stuff, now, but I think that ... People in the chat actually can either say Leary's completely off, or close, or what ...? Enrico Palmerino: This is why I'm saying I don't think ... We're making as if Botkeeper's ... One, granted, yes, [00:50:30] we're treading new territory by immersing AI into the accounting industry. This is something that ... AI's entered other industries first, and now it's coming to the accounting industry. When you enter an industry, you follow suit with the other tech companies in the industry. To your point, like you just said, there are companies that do a piece of the bookkeeping equation that refer to that as bookkeeping. We're saying Botkeeper ... Because Botkeeper doesn't fully [00:51:00] automate every aspect of bookkeeping, that we can't say that we have bots, but we do. We have machine learning; we have AI; we have natural language processing; we have robotic process automation; we have workflows, decision trees, OCR. All of those tools are tech that automates tasks. You can call any cluster of them a bot, because it's performing some of the action. This company, for instance ... Not to single you out, David, but you're [00:51:30] here on this discussion with me ... That company that you represent, they don't post on their site - front, bold letters - they don't ... I don't even know if they have a Facebook page for their offshore labor [cross talk] David Leary: Every webinar I do, it's, "Hey, there's human verification that also takes place." We've always said it [cross talk] Enrico Palmerino: -on the same page here. I totally agree, and I just don't ... Why is it being made such a crazy point that, because ... If you go to [00:52:00] our website, we say 'humans.' We say, "A Botkeeper is a combination of skilled human accountants plus machine learning, and AI." We published that in most of the articles we published. We don't ever claim that this- David Leary: I think that's something you talked about on the phone, like everybody misses that part, right? Blake Oliver: Well, but wait [cross talk] Hold on, hold on, hold on. I would like to read one of Enrico's public statements and ask Enrico to clarify this. I highlighted this in my post. "The easiest way to think [00:52:30] of Botkeeper is as a robot bookkeeper. It replaces the need for a full-time bookkeeper in your company. It will pay your bills, invoice your clients, reconcile your bank accounts, produce your financial reports, classify expenses, and revenue, administer bill-pay workflows, and payroll, and do any calculations needed for this. It can be the last bookkeeper you'll ever need to hire, and very quickly become your entire accounting department. Since creating the Botkeeper, we've never had to hire a bookkeeper." Who are the people in The Philippines? What are they [00:53:00] doing when they validate those transactions? Isn't that bookkeeping [cross talk] Enrico Palmerino: We weren't saying since ... Since we created Botkeeper, WE haven't had to hire a bookkeeper. We were saying that Botkeeper, as a thing, as a company, is a substitute for a bookkeeper. Where you would normally have a human bookkeeper in your office doing bookkeeping for your business, you could instead not have that person sitting next to you, sign a contract with Botkeeper to do that work, and you wouldn't [00:53:30] need that person here, and that it is a robot bookkeeper, a cyborg, however you wanna define it. What we're trying to do in this ... Anytime you try to convey a new technology, or a new system that leverages various aspects of technology, and is complex ... I gave the background of what's actually happening, when Justin ran that thing. More often times, a company communicates not ... Communicating all the complexities, and diving [00:54:00] deep on the tech isn't gonna do anyone any good. It's like trying to describe all the nuances to what makes a phone operate. What you care about is that you can take photos on it. It has an alarm clock. You can make calls. A phone has evolved so much, since it began seven years ago, or since the first more modern cellphone [cross talk] Blake Oliver: Yeah, but part of the reason I use an iPhone is because I have guarantees from Apple that my data [00:54:30] stays on my iPhone, and that it doesn't go overseas, Enrico. Enrico Palmerino: What about people who use Androids? Blake Oliver: That's why I don't use one, because I can't trust it, all right? I'm kind of flabbergasted ... Can you not see how an outsider, who doesn't know anything about bookkeeping [cross talk] Enrico Palmerino: -wait, aren't you on Facebook? Blake Oliver: What's that? Enrico Palmerino: Aren't you on Facebook. Blake Oliver: Oh, yeah, I know [cross talk] Enrico Palmerino: -you wanna say you use technology that guarantees your data doesn't- [00:55:00] can't get hacked into, and get out ... You use technology every day- Blake Oliver: I know, and Facebook is terrible, and it's awful that we have to use it to get [cross talk] to people. Enrico Palmerino: But you use it, so the argument that you only use Apple, because it only does this is ... You use tools all the time that don't ... Who's to prevent [inaudible] from getting that? Blake Oliver: Here's the thing is that's why there's been all this reporting on Facebook. We should know what is happening with our data, and where it's going. [00:55:30] That was the intent, I think, of the ... Well, I don't think. That was my intent in all of this is not to- Enrico Palmerino: You know they build iPhones overseas, right? Blake Oliver: Right and guess what. Here's ... Let's say let's say Apple made those iPhones, and they said they were built in the US, but really, they're built in China. I think we would all ... That wouldn't be fair, would it? Enrico Palmerino: We've never claimed Botkeeper as a US-only company. We're a global company with global clients [cross talk] Blake Oliver: Where does it ... In your terms of service on your website, it says that all the data [00:56:00] is processed in the United States. Did you know that? Enrico Palmerino: Refers to computing, and all the data sits, and lives on US servers [cross talk] by federally regulated ... Our legal team had ... We went through an extensive process, in terms of how we decided to build our servers. Actually, quite frankly, managing servers that were in the US versus managing overseas is just easier. We went through extensive legal review on how we did these contracts to confirm, and basically use the same terminology that most [00:56:30] tech companies use that the data is living on servers here. Blake Oliver: Is validating not a form of processing? Enrico Palmerino: This comes back to, like, David saying [cross talk] it's a bookkeeper, an accountant, a CPA ... We're getting into semantics. I understand ... Don't get me wrong, I totally understand the point here. We want to get better. I want to help you get better. I need the help of the accounting industry. You bringing this [00:57:00] to light is real ... We didn't realize this was a problem. People, our partners, clients weren't coming to us, and saying, "Hey, we found out you had a PH office. This is really frustrating. We didn't ..." because they already are using ... Most accountants, especially cloud accountants, you're already using technology that has offshore offices, and that is, in one instance, actually sending, truly sending the data overseas. Our offices are actually our offices. We don't go to Mechanical Turk, or something. [00:57:30] David Leary: I think part the part that really ... If I look back at the last 12 months or so, because you made a big splash at all of the accounting conferences; you're doing talks on stage, et cetera, et cetera. A) The name is genius. It's a perfect name. You have good marketing. Also, a little bit, I think, the confusion is -and this is even for me - because I have people ask me, "Hey, they're an app; are they an app?" because everybody asks me about the apps in our space. Then, I finally figured out, no, they're just an accounting [00:58:00] firm with tech, or a bookkeeping firm with tech. As soon as you comprehend that, you think about what you're doing differently. Outside of that, I came up with that conclusion. I don't think that's the average conclusion people come up with, because really, it's always like the ' magic bots," or ... I think even Jody Padar was like, "Oh, it's Enrico's magic brain. You put 'em in the bots, and it does everything magically. The bots are doing my whole company now." You came off, really ... You've come off as being this miracle [00:58:30] technology, and you're really a bookkeeping firm that has some serious cool tech, and is super highly efficient, but you've just never came off as this super-crazy-efficient bookkeeping firm, with a suite of services, like you kinda said in the very beginning of this call. Enrico Palmerino: I think that's because what we wanted, when ... Like you said, we're very different from a traditional cloud accounting firm, or from a traditional bookkeeping firm. When you're doing your marketing, you market the differences. What makes Botkeeper different [00:59:00] from another accounting firm? Well, we've got 30 engineers who have built some really cool tech to allow us to do more with less resources; and not just do more, but make it more accurate, give you more reporting, more analytics, do it faster. Those are the things, when we do a lot of our messaging ... Yes, we provide bookkeeping as a service, but why would you use us over someone else? Because of years of tech that we've built, and deployed, and that's what makes us ... [00:59:30] Blake Oliver: Having come from a firm ... I came from Armanino, Top 25 accounting firm, before I joined my current company. We were competing against guys like you, Enrico. We were only using accountants in the United States, and it was very expensive, as you know, to do bookkeeping in that manner, but it was a selling point for us. We could say that your data is in the US, and it's being processed by Armanino employees in the United States. I [01:00:00] don't have to compete against you anymore, which is kind of nice, because it was difficult [cross talk] Enrico Palmerino: -we'd rather be working with you [cross talk]. Blake Oliver: There are other similar services, I should say, right? What was unfair about it, what struck me as very unfair, and I think other folks in the chat will agree is unfair, is that you've got this amazing marketing, where it says that it's all automated, and it's bots, and it's not talking about the humans; it's not talking about The Philippines, and it puts them at a disadvantage. [01:00:30] How do you respond to that? Enrico Palmerino: Blake, have you seen the first page of our site? Just read what the definition of Botkeeper is. What is Botkeeper? Blake Oliver: You say it's a combination of artificial intelligence and humans. Enrico Palmerino: It's humans first. We actually even, in the way that we describe it, we say, "skilled human accountants," and AI. Blake Oliver: Right, but you don't talk about ... This is where I get to ... I said there's two points here, and we shouldn't get confused between them. There's the fact of human versus machine, which is less important to me than the failure to [01:01:00] disclose offshoring. There's a lot of people in this country right now who are concerned about job losses, due to automation. It's kind of like you're taking US bookkeeping jobs, and you're shipping them overseas under the guise of AI. Enrico Palmerino: Oh, come on. That's really harsh. I mean, we employ ... I wanna say Botkeeper probably employs about 100 people now. We've hired 30 people here in the US, in the last 30 days. We're [01:01:30] hiring a lot of brilliant people, and I think ... My understanding of the accounting market, and any time I've done these talks, I said, "What are the biggest challenges everyone's facing?" It's they can't find good bookkeepers, so I don't think it's a ... I think Botkeeper, if anything, is helping augment this lack of supply. Hiring, maintaining ... You even told me yourself, you said, "Look, I would not be able to hire ... I could not ... The reason I got out my business - it was not gonna be profitable, and scalable for me to try to find another bookkeeper [01:02:00] like I had. I just wasn't gonna be able to repeat that." The reason that that's the case is because there's not enough of them. There's almost accounting firms from the baby boomer generation than there are new accountants entering the industry to take them over. I think we're trying to help bridge a gap. We're certainly not trying to cause job disruption. There are a lot of people that are employed by Botkeeper that have built lives, and families with Botkeeper. We take care of our staff. We give them benefits. [01:02:30] Please don't make us out to be this like evil company trying to put everyone out of jobs, because you could ... Why isn't the platforms that we use every day to do the accounting in an evil company that's putting us out of jobs, because they have a thousand people in India, and people over in other countries? [cross talk] David Leary: -people digging holes with shovels- Enrico Palmerino: C'mon, man. This is like mountain out of a molehill. Blake Oliver: Well, so, the thing [01:03:00] that I would - again, I keep coming back to - is simply disclosure, and I'm really, really happy to hear that you guys are gonna be more upfront about this in the future. That satisfies ... I wish you all ... I really do, I wish your company success, and I hope that you achieve greatness, and really do build a tool that can automate a lot of this manual work. The thing, to me, that's [01:03:30] been disappointing has been the appearance of, "We've built this magical thing," when, in fact, you haven't. I know that's a big tech thing; that it's very common in the tech world to prop yourself up with really clever marketing, and [cross talk] Enrico Palmerino: -we built a magical thing, when we did it, but we just ... We started this whole thing with doing a demo where, if there was thousands of transactions being processed, it would normally take a person a lot of time to do even just [01:04:00] that one piece of the equation. Botkeeper has magic that does it. It's not actually magic, obviously. It's a combination of technologies that we built, but we're touting the tech that we've built, because we're very proud of it, and we have people that put a lot of time, all-nighters, into doing it. We should be excited, and proud of it. Blake Oliver: Well, yeah, but again, Enrico, don't pat yourself on the back too much, because Amanda Aguilar said ... I don't know if it was in this thread, or on Twitter, she said, "Hey, I can automate 90 percent of the bookkeeping just setting [01:04:30] up some bankrolls." Justin DoBosh: [inaudible] yes, there's a combination of humans, and accountants. I started back at Botkeeper in November 2017, so I know every part of the Botkeeper system. I've built most of it. You can ask anyone at Botkeeper, I sleep about eight hours a week. That's because I care so strongly about our clients, that they use our software, that it's like ... You can ask most of our clients ... Some of them email me directly, at 2:00, or 3:00 a.m., [01:05:00] and I respond back to them, and they we make sure ... When you say, Hey, you guys just made a bunch of magic," no. We have 30 engineers, who give up their weekends, give up ... Even our Philippines office, because we have engineers over there. Some of 'em commute five hours a day to come work. To kinda just say we're not building anything, that's a pretty big backhanded compliment to all the work we've done here. I just want you to know, everything here ... I would just appreciate it, if you could be [01:05:30] a little bit nicer here [cross talk] Blake Oliver: I'm sorry, Justin. I mean you, and the engineers no disrespect. Just from an outsider's perspective, we came here to see a demo, and so far, the only thing I've actually seen is you guys code some transactions from a bank feed. I'm just saying where is it? Show me the bots. Enrico Palmerino: If we were selling something, if we were creating a hyperbole on what we've built ... We've all [01:06:00] seen what happens when a startup way oversells, what is actually done. Someone buys it, and they realize it can't do, or deliver the thing I bought it for, they leave, or they leave bad reviews. You started the conversation out with being like- David Leary: That's true. Enrico Palmerino: "You must've done something [cross talk] pretty special, because you've got clients that love you; you have partners that love you. You guys have built ..." We've never claimed that ... We're, [01:06:30] like I said, a cyborg like. This thing is creating better bookkeeping. If there were no humans, we wouldn't have a team of people. The goal is to make better bookkeeping; to do it more efficiently, to allow ... I think the thing that is scary about the industry is the fact that some of these platforms that we all leverage to do accounting in are potentially going to be doing bookkeeping. We built [01:07:00] a platform that allows our partners to offer, still offer bookkeeping services at very competitive rates; make some really nice margin on them, and get a better result than they would if they were trying to hire, and manage a team, and that's because ... It's that combination of tech, and human, and even just how that ... There's an intricacy there, like how the tech interacts with the human; how the human leverages the tech ... That's not easy to do. I think we've done a really, really good job at it, and that's why we've got a lot [01:07:30] of people that love our product. Blake Oliver: That's great, and I give you all the credit for that. I think, to kind of wrap this up, probably, we should get toward the end, right, David? [cross talk] David Leary: -wrap this up, right. Blake Oliver: I'm sure Enrico's got incredible stamina. I have to say, Enrico, your public speaking ability is incredible, and your ability to answer every question is just amazing. I don't think I could do it. Thank you for giving us all this time. I guess the way I would summarize this is sort of like we've almost got a culture clash here. [01:08:00] No, we do have a culture clash here. It's the tech culture, and the accountant culture. It's clear to me that, in the tech world, which I am now a part of, as a CPA, and learning all about, it's clear to me that there is this mentality of, "Let's just make it work. Let's make it happen, and let's get it done ... We're not gonna think too much about how we do it." In the accounting world, we think a lot about how we do it, and the [01:08:30] ethics of how we do it, and how we disclose it, and how we talk about it to our clients. It's not enough for me, if I'm a CFO, or a controller, to go and print a bunch of financial statements, and say, "That's it. Doesn't matter how that happened; doesn't matter how those numbers got put together. The numbers are good." No. How they get put together matters. I would just say to anyone who is watching this, in the tech space, it's really important, I think, to think about how you are doing it, [01:09:00] and how you are disclosing it, and whether or not you are being fully upfront, and transparent about it. Enrico Palmerino: I think I agree with you. It's two different cultures - technology entering a different industry - and we need to ... To solve this, and to make it work, it can't be one-sided. We need the accounting industry to ask the questions that they care about. If "Do you have a team overseas doing anything?" is an important question, then they should be asking it of every technology [01:09:30] company that they want to leverage. The same way that we're learning that the accounting industry cares about those things, let's start being proactive in disclosing them [cross talk] Let's help each other make this industry better. Blake Oliver: You know who we can all hate on right now is the media, because I think that they kind of failed here, in that they didn't ask you any questions. Enrico Palmerino: The [01:10:00] challenge with the media, too, is media writes content about their interpretation of something. Botkeeper emerges; someone reads about Botkeeper, writes their opinions about what Botkeeper is, or how it works, or how it acts, and then publishes it as if it's totally fact. To this point, there is a publication out there ... I remember when it came out. It said, "Botkeeper is a chat bot for accounting," and I-. Blake Oliver: Yeah, TechCrunch wrote that. Enrico Palmerino: That was the first time I [01:10:30] had heard that. I was like, "Okay, you think we're a chat ... We're not a chat bot. We don't even have chat bots internally," at that time. We leverage some chat bots now, but ... The problem is they published that as if it was fact, and never did this, like what you and I are now doing, here, where you're asking me the detailed question; I'm giving you a detailed answer. All too often, the media, to your point, shoots from the hip first, and then asks questions later. I [01:11:00] think they need to do a lot more of this, where you can ask questions, they can answer them; then they can actually document the fact. There's an issue out there with fake news, and every snowstorm these days is a Snowageddon, right? It's all being blown out of proportion. I'm looking forward to, like I said, accountants, and tech companies working together. We're trying to make the industry better. We want to work with you. Help us. David Leary: Yeah, so, I [01:11:30] think we should start wrapping this up ... We don't wanna just go on into people's happy hours now, on a Friday night. Thank you for joining us, Enrico. Did you ever more thing, Blake? Blake Oliver: No, I'm- David Leary: I see that you're ready to dig in again. Blake Oliver: No, no, I was giving you a thumbs up, David. David Leary: Okay. I honestly think, from a perspective of things just haven't been clear, and by taking the time, being a forum like this, and the people who ... Obviously, people felt it was important. I think some 40 people or so attended to learn more about [01:12:00] this. Now there's a place for people to go, who want to learn more about Botkeeper, can go to this video, and learn more. I think it's very balanced that kinda came out of this. Thanks for coming. Justin, thank you for doing the demo you did. Blake, I think that's it. You wanna wrap up, and call it ...? Blake Oliver: That's it. David Leary: All right. I think we'll kill it. All right, bye, everybody. Blake Oliver: Bye. Enrico Palmerino: Take care. Justin DoBosh: Thanks. Bye.