Artificial Intelligence Part 1

There may be errors in spelling, grammar, and accuracy in this machine-generated transcript.

Hector Garcia: Welcome to the unofficial QuickBooks accountants podcast. I am joined by my good friend Alicia Cat Pollock, the original, the one and only Qbo Rockstar CEO and founder of Royal White Solutions.

Alicia Katz Pollock: And I have the privilege of collaborating with Hector Garcia, CPA, the founder of Right Tool for QuickBooks.

Hector Garcia: In this episode of the unofficial QuickBooks accountants podcast, we're [00:00:30] going to talk about artificial intelligence. Alicia, thoughts on this topic?

Alicia Katz Pollock: Well, it's the topic that everybody is talking about, and it's already showing up in all areas of our life. And I know that it's been really helpful for me for a lot of my writing. And so I'm really interested in hearing what you have to say about what where we're at and where we're going. Yeah.

Hector Garcia: So artificial intelligence has always been discussed Discuss mostly [00:01:00] in a fiction scenario, like in movies and stories and that sort of thing. And there was some semblances of artificial intelligence life, let's call it that, um, prior to December of 2022, but pretty much ChatGPT changed the game, right? So ChatGPT became a word, a household name. When people think of artificial intelligence, they might think of ChatGPT first, although there's a lot more other options out there. You know, there's cloud AI, there's Google Gemini, [00:01:30] there's meta AI. Uh, Twitter or X is getting into the into the artificial intelligence game. The reality is that this this is not owned by one brand anymore. But I would say December 20th 22nd January 2023 when ChatGPT blew up, it was all ChatGPT this ChatGPT that. So, um, one of the things that I made the conscious decision of is, look, ChatGPT will affect the accounting Bookkeeping. It will. It [00:02:00] will affect our profession in one way, shape or form. I'm not sure what that would look like, but then I'll start creating some content around that. So I went into YouTube and I posted a couple of videos of me experimenting with ChatGPT. So some of the core things I started doing is what if I post a PDF bank statement and say, could you convert this into like a clean table that I can import into QuickBooks? First attempts of that not so good. Today I do it. You get 90% of the data clean organized. Okay. [00:02:30] So it's, you know, only two years in a year and a half in. And they made so much progress where they will be a point where you getting a QuickBooks file, getting access to QuickBooks, getting a bank statement will all be the same, basically because you throw it into ChatGPT and ChatGPT can convert it into anything that you want it to be.

Hector Garcia: Um, some of the other things I started experimenting with is grabbing maybe a CSV file of a bunch of credit card transactions downloaded maybe a month worth. [00:03:00] Throw it into ChatGPT and ask it, hey, how would you categorize these expenses for a small business? Maybe using the chart of accounts or the list of expense categories in a schedule C form for, you know, for really small businesses and ChatGPT would, I would say in maybe 40% accurate, accurate, it would put Chipotle in meals and entertainment. It would put American Airlines in travel, and it will put, you know, Netflix on there, you know, other or something like that. And the reality [00:03:30] is that the nuance that we add that bookkeepers add into any of these things, into converting files, into categorizing transactions, includes context that ChatGPT might not have, which is and the classic example that I use or that I used to use when I used to teach QuickBooks every single weekend of my life for five years in my training room in Miami was, if you see an expense for water. Okay. If, depending [00:04:00] on the business, water could be a different category. For example, a laundromat. Water is the most expensive direct cost. That would be a cost of goods sold. But if it's an office. Right. Water could be, you know, office supplies because you're just buying water for, you know, for for your employees or whatever.

Hector Garcia: So that piece of context, ChatGPT is not that good at yet. And the point is that ChatGPT can do or any AIS can do a really good job at grabbing [00:04:30] a core knowledge source, knowledge base and a prompt and the variable. And we can deep dive into these three things. These three elements are really important, and then try to produce something that the accountant or the bookkeeper used to do manually, or by thinking about it prior to it being done. So this is the core of what I've been thinking about ever since ChatGPT came out. And all of the other competitors to ChatGPT as to where, where, [00:05:00] where or when will be the point where this stuff will actually take over some of the stuff that we used to do manually and eventually, uh, take away some hours of work from us and eventually take away a staff member or a head from our team because because what we used to be able to do with with five people you can do now with four people and then with three people with the aid of ChatGPT. And this is kind of like where I think things are going. So that's the background I want to give you in terms of like how [00:05:30] I feel. Um, artificial intelligence is going to start seeping into our profession and start affecting what we do. And then maybe we can talk about like specifically how this works in QuickBooks. But Alicia, any thoughts on this?

Alicia Katz Pollock: You really touched on my biggest puzzle is that at the moment, you can't really take people out of the equation because, you know, if you have a receipt for Starbucks, how do you know if it's the owner getting coffee on their way to a job, or them sitting down and having [00:06:00] a 1 to 1 for Benny? You know, you there's no way of knowing without having some additional information. So, you know, that's why I, you know, look forward to still having humans in the equation for a really, really long time. But being able to do some of that more monotonous work of categorizing the transactions, you know, we've already had the bank feed in QuickBooks online doing a lot of suggestions for years. And it gets better the more you use [00:06:30] it. And every time I'm working with a client, they're like, these are all wrong. I'm like, actually, you're in charge of the bank feed. It's not telling you what to do. You're telling it what to do and you have to train it. So that's the key is us training the AI so that it does get more accurate over time. And, you know, it takes a couple months to train your AI in QuickBooks online, but it's going to take a couple years to train it for all of the different contingencies inside accounting and bookkeeping.

Hector Garcia: Yeah. [00:07:00] And one thing I want to point out is prior to like ChatGPT and all these AI, um, coming out and becoming mainstream, becoming something, you know, any it's very, very democratized. Anybody can play with AI. Quote unquote, especially with the large language models. Prior to this, I thought that what Intuit was doing with bank feeds, kind of the way you're framing it, was AI. But I still believe that there's no such thing as AI in bank feeds in QuickBooks online right [00:07:30] now. What I think bank feeds is it's just a numbers game, right? So if the majority of businesses put Chipotle as meals and entertainment, then it would suggest meals and entertainment. It doesn't take context into this is the example I'm giving about the water and the laundromat versus the office. What I'm saying is, I don't think bank feeds takes into consideration the context of the other information that it knows about [00:08:00] the company. Now, you said you got to train the AI. Okay. That's different. Okay. So what? Training the AI means in bank feeds is. Bank feeds will keep note of how you categorize a vendor you used in the past, and it will override the general knowledge that the majority of the 6 million businesses that use QuickBooks picked Chipotle as meals and entertainment. But you in the last three times that Chipotle came in, you put it as a personal expense. Therefore it will override, [00:08:30] you know, the general knowledge information with what you've done in the past. So I think that's what you mean by training the AI. And there's also rules that actually it's one step above that says hey explicitly, if it contains these words, you're going to follow this directions. If it doesn't contain these words, then you're going to follow last behavior. And if there isn't a pattern of last behavior, then you're going to follow the general crowd. The crowd sourced information that Intuit has from all businesses doing bank feeds. So [00:09:00] I get that. Is that do you agree with that?

Alicia Katz Pollock: Yeah. That's not artificial intelligence. That's just intelligence. That's programing. Rules are essentially user programing of the software.

Hector Garcia: Right. So now let's talk about where it will go when you actually add quote artificial intelligence into it. Right. So we talked about bank feeds right now being something that's programmatically set up in terms of rules and priorities. Right. So we have history rules and priorities. You follow it when you add artificial [00:09:30] intelligence you're going to add context. So maybe right, uh, QuickBooks will look at an Office Depot expense that is $2,500 and say this is going to be a fixed asset, but it looks at the revenue of the company, and the company is $10 million in annual revenue and goes to $10 million in annual revenue. It's probably not a fixed asset. It's probably an office expense because of the relationship between the transaction and the overall [00:10:00] set of information and context that it has about that company. At that point. When it does that, that will be artificial intelligence, right? Because it will it will start. Let's just I'm going to use the word wrong word here. It will start thinking or discerning multiple aspects of it right into it. Or QuickBooks ask you what industry you're in. So when it takes into context industry, when it takes into context business size, when it takes into context, maybe the fact that you've never dealt with that vendor before takes into [00:10:30] context the fact that you've done, done used that vendor before, like all these different pieces of context, like, for example, let's say you go to a Home Depot and Home Depot tells you the store number, it will go out and search the store number and the physical location.

Hector Garcia: And then maybe that store number is close to a job that you're doing. It will maybe suggest, you know, that project or customer job. That's the stuff that humans do now, like we do. That's all the thinking that's happening in real time, all the context that we're taking in real time. Right? Like when I do, um, [00:11:00] you know, when I do accounting for a client and I see a restaurant, I don't even think about it. I put it on their meals and entertainment or meals or business meals. When I see a restaurant with a French name and a French address, I put it into personal because I assume that has to do with the French. The France trip that I know my customer took because we talked about it last month, and that is the the part that is very difficult to automate, you know, through artificial intelligence is [00:11:30] that context piece. But this is where this is where statistics come into play. There will be a point in the very, very near future. And I did a video on this on YouTube, and I'll put a link to this in the show notes, where the user will have the choice not to pick any categories, they will just accept the first thing that's, that's that's statistically QuickBooks would suggest should be the category to every single transaction and literally [00:12:00] prepare a PNL and a balance sheet simply by connecting your banks and using, let's call it, a combination of patterns and AI to build a PNL and balance sheet.

Hector Garcia: Now, every accountant in the world will look at a PNL and balance sheet, and it will be able to dissect it and point out everything that's wrong with it. We will. We will be able to. And trust me, there will be a coalition of bookkeepers and accountants that will attack this automatic PNL and balance sheet creation. But [00:12:30] after you go through the the, you know, butchering of the, you know, the, you know, the splitting apart and dissecting of this thing. If the real variance in net income or taxable income is no more than 5%, guess what? Customers will not care. Customers will not care. So our role, our role is to make them care. And telling them that, oh, it's this category and not that category. It's not going to convince any business owner to care. [00:13:00] Like our role. And I don't even I don't even know what that looks like. But our role is to humanize AI and say, hey, AI has faults too, just like us.

Alicia Katz Pollock: Okay, this is where my OCD kicks in because I care and I want my clients to care, and my clients care that I care. So, I mean, that's one of the things that makes us valuable as financial professionals is the fact that we do care. And then the business owner realizes why they should care. [00:13:30] You know, there's growth there. So I'm not sure where where AI is going to be able to create that relationship and dynamic.

Hector Garcia: Yeah. Funny that you mentioned that. Um, one of my favorite books is called Exactly What to Say by Phil Jones. And we a different episode. You kind of brought it up and I completely. If you wouldn't have brought it up, I wouldn't have thought about this. But he says in that book that our job is to care about what the people that you care about, care about. To care about what? The people you care [00:14:00] about. Care about. Yeah. And this is the thing, just by extension, because we care doesn't mean our customers care. So the reality is, and this is the problem that our profession has from a branding perspective. The reality is, is that 90% of the motivation that comes behind hiring an accountant, hiring a bookkeeper to make sure their QuickBooks on their books are accurate is to make their taxes right. Right. To make their books right. So the taxes are right [00:14:30] so they don't overpay taxes or they don't pay a penalty. So there needs to be a really important shift to that. Because if we continue to depend on the motivation behind hiring an accountant, being that the what they care about, quote unquote, is having their books right for taxes. And if artificial intelligence gets to the point where it can be, again, within only a 5% variance of error, and even if the IRS penalty is four grand or something for that little variance [00:15:00] of error, and that's cheaper than hiring an accountant, guess what? The market is going to shift and Intuit it's going to be in a very interesting position.

Hector Garcia: And I believe and this is very much speculation and conspiracy theory and all that stuff, we can add all we can stack up, all these negative things that are going to be what I'm about to say. But I believe that QuickBooks Live what Intuit is doing with QuickBooks Live is a means to an end. I don't think QuickBooks wants to be in the service business. I think the most profitable industry in the world is software. And interestingly [00:15:30] enough, second is accounting. So. So Intuit is really not trying to figure out how to make services profitable when software is already profitable. What QuickBooks Live is doing is they have an army of bookkeepers working for Intuit, giving them insight as to what is it that the humans do so they can take that knowledge, plus the work that we're doing by, quote, training the bank feed and all that stuff and build more automations behind it. Now, you could, you could, you could I could frame that as [00:16:00] super macabre. And they're trying to like, take over, you know, accounting, the accounting industry. Or I can frame it as they're being innovative and they're trying to add more value to their customers depending on the regardless of the framing. Our job is to rebrand the profession as one that matters, that our work matters, the nuance matters, and it has to be much more beyond, you know, the variance of a, quote, accurate PNL and balance sheet versus what QuickBooks will be able to automate in the very near future.

Alicia Katz Pollock: Uh, [00:16:30] just as an aside, it was it's only been a couple of months since we had that whole brouhaha with the announcement of Qbi Live Expert assist, and everybody was like, oh my God, you're competing with me and you're gonna put me out of a job. And so far I haven't heard of a single person who has been put out of a job by QuickBooks Live expert assist, and we know that expert assists its fundamental purpose. Like you're saying is, on the one hand, on the surface, it's a paid support line. On [00:17:00] the the the deeper level, it's an opportunity for Intuit to see hands on what their clients are doing, how they're doing it, and gain that context that they can, you know, figure out what to emphasize in the software or the usability of the software or the programing behind it, so that they can automate some more things in the background. And so we know that the purpose of a live expert assist at the deepest level is more business intelligence.

Hector Garcia: Yeah. [00:17:30] From from a micro I could totally sympathize with someone from a micro level looking at Intuit offering accounting services or elevated QuickBooks support as being a competition to what we have been doing in the past 25 years as ProAdvisor. Okay, I can totally sympathize with someone seeing it in that perspective and in again, in a micro short term, it could feel like a direct competition, but in a macro level, it's really good news that that the all powerful Intuit has recognized that you need humans to [00:18:00] do this job correctly. Okay, not just that you need humans to bridge the gap between the promise of the software and the user and the user input. So QuickBooks Live exists because the need for human exists. And then and the reason why they did it in-house, other than the conspiracy theory that I have, that you're just trying to learn how to figure out how to train the AI so they can automate it and get rid of all the accountants. Other than that, the reality is that QuickBooks [00:18:30] is satisfaction. The satisfaction level that people have towards Intuit QuickBooks has not improved in the last couple of years, and Alicia and I will tell you, 90% of that is user error. Like most people that complain about QuickBooks is you go to Reddit, you go to the Facebook groups. 90% of complaints are user errors. They're misunderstanding of what your role is as a human. And the role of the software is and but it's very much fueled [00:19:00] by Intuit's misleading marketing that you don't need an accountant. Okay, now you can say Intuit hasn't said doesn't doesn't say that right now, but true. They don't say you don't need an accountant, but has their accountant relationship actually gotten better over the years? You know, have they actually increased the benefits to accountants or decrease the benefits to accountant? Have they raised the price on their accountant? So there's so many things that feel like posturing towards we can bypass the accountant either [00:19:30] with QuickBooks Live or eventually with a solution that can do it all, and it feels that's the direction that they're going to.

Hector Garcia: And as much as I hate it for my people I love it for my stock because Intuit is going to, you know, continue to go up in stock price because again, if for years they have published that accounting is number one or number two, most profitable industries talk about accounting services. And software has for years always been it's only natural for software and accounting to sort of go [00:20:00] hand in hand. But they control the entire ecosystem. So um, A ProAdvisor which still exists, support for accountants, which still exist. We did an episode of what Intuit does. Right. And we said accounting, education and accounting partnerships is what Intuit has been doing right for a very long time. But right now they need to be very, very, very careful on how to thread the needle of these two realities that are happening, happening at the same time. Reality number one, business owners need accountants, business owners trust, [00:20:30] accountants. Reality number two, business owners will expect their software to do everything automatically and bypass the accountant. So these are two realities that are happening at the same time and our job. You, as a listener of this podcast and myself and the entire industry needs to move towards rebranding our profession to to set our place as the human advisor alongside the AI automation. Mhm.

Alicia Katz Pollock: Yeah. Instead of people being the people who are actually typing in the information, were the people who are analyzing [00:21:00] it and verifying it and double checking it instead of actually having to spend our time typing in. Here's the date, here's the payee, here's the account, here's the backup. You know, it totally transforms what we're doing. And I think that's far more interesting to be able to look over the data and talk about the data than it is to sit spending my time actually manually creating transactions. So I'm really, really, really glad that we are where we are and [00:21:30] look forward to going where we are going.

Hector Garcia: It's just difficult to digest for the person that has made 90% of their income from manually entering data in the last 20 years is for that person is very hard. So we have to sympathize with those people and like think our job as educators. Alicia and I, we kind of have to make the commitment that our job is to get the people up to par, you know, bridge that gap. This is why I create those videos about ChatGPT, not [00:22:00] because ChatGPT is going to solve the problem. It's because this is a good illustration of where things are moving, and this stuff cannot catch you by surprise. This is really the most important thing about artificial intelligence is. One thing that you know for sure is that it's going to get better every single month, and one thing that you know for sure is you can either keep up with it or let it catch you by surprise, and when it catches you by surprise, you might be so many steps behind that you won't be able to catch up. And that's a really important piece. Piece [00:22:30] of this.

Alicia Katz Pollock: Yeah, no. And I'm actually going to suggest that this is a really good time. If you haven't started exploring I or not? This is a really good time to start taking a look at. For example, how do you use prompts? How do you actually ask questions accurately to get the information that you need? And of course, it's all going to change, but this is the time for everybody to start doing that. I've been following Ashley Francis and her kitchen table AI, and she's [00:23:00] teaching people how to relate to the computers so that you have an actual relationship with your AI. And I really feel strongly that it's not something that you can ignore. But even if you just dip your toes into and I see it everywhere, like even in my email program, there's a little AI prompt down on the bottom where it will take what you write and rephrase it for you and make you sound just that much more professional. And for a lot [00:23:30] of people, that is, you know, we used to complain complaint about spell checkers back in the day. Like knowing how to spell. Used to be when I was growing up. Really important. And then spell checkers became built into Microsoft Word, and then all of a sudden it would just either correct it for you, or at least let you know that it was spelled wrong. And so needing to know how to spell became a skill that was helpful but no [00:24:00] longer required. And that same kind of evolution is what's happening now. What is important for people to know and do, versus what can be taken care of for us so that we can put that brain power to other things other than knowing I before E, except after C, or when sounded like A, as in neighbor and weigh.

Hector Garcia: You know, um, that's really interesting that you say that because one of the one great illustration of someone that's probably [00:24:30] regarded as a pretty sharp individual like myself, I'm a horrible speller, and the reason for that is because I don't need to be a good speller, because as I'm typing stuff, this stuff is being fixed for me. So I wonder if we can draw any analogies from this specific concept. You're saying that, you know, 20 years ago being, you know, being a good speller was a very good thing. And today is to day and age. You could be successful being a bad speller where 20 years ago you probably couldn't. Um, one of the interesting pieces about this is in accounting. Like [00:25:00] there's like spelling is a fundamental part of written communication, right? We agree on. We agree on that. And knowing which is the debit and the credit is a fundamental part of accounting. And knowing how to categorize expenses within the context is a fundamental part of accounting. I wonder if I will do to categorization or debit and credits to what spell checkers did to, to, to spelling, and will that fundamentally change the stuff that [00:25:30] we care about as fundamental And just completely ignore it because the eye is sort of like doing it. Like if we're doing a, we're doing a journal entry and we put something as debit, I will automatically put it as a credit because it knows that that's a natural credit account or whatever. So I wonder if that's the true impact won't be on that. It will just be on that, you know, doing things on the fly for you that you don't even like. Think about the importance of it because it's out of out of correcting, you know, or spell checking your accounting work.

Alicia Katz Pollock: That's exactly where I was going with this. I [00:26:00] mean, QuickBooks, even though QuickBooks itself isn't I, it has already separated the need to being able to do debits and credits in your head. You still need to understand them, but you don't need to be able to instantly know this is a debit and this is a credit to the point that, like I have a chart above my own computer that with with the with it, so that when I'm having a zoom with somebody and I'm talking about debits and credits, my eye is going up to my chart to make sure I'm getting it right before [00:26:30] I, before I actually say it So that's one of the things that will change. You know, accounts absolutely, positively need to understand the debits and credits behind the scenes. But being able to just have to say it is already changing. There's millions of people using QuickBooks online who have never even heard the term debit and credit. Business owners don't know that. And a lot of the bookkeepers who are starting with bookkeeping buy through the software [00:27:00] side and not from the accounting education side, you know, they've heard the terms, but they don't need to use them on a regular basis. And so that's a perfect example of information that's not going to be as required for success as it is right now.

Hector Garcia: Absolutely. And since this is a QuickBooks podcast on official QuickBooks accountants podcast, let's talk about what I actually see inside QuickBooks online. That's quote, artificial intelligence. So [00:27:30] Intuit announced September of last year, September 2023, that they were coming up with this concept called Intuit Assist, and they had this financial language model. They had all sorts of like really funky framing and labeling that makes it makes made all of us feel that Intuit is an AI company now. Okay, so a year and a half a year later. So let's say a year later, what do we see? So when your when the Intuit assist is [00:28:00] enabled in your QuickBooks file, you would see a big button in the top that says Business feed. So they call it business feed not Intuit Assist. But Intuit Assist was like the codename that they gave it back in September when they announced it. So when you click on Business Feed, you see sort of like an invoice creation assistant. So it says, you know, forward a conversation with your customers from your QuickBooks admin email to a private email address. So the concept behind this is [00:28:30] they give you this funky email, okay, that maybe a customer says, hey, can you ship me another widget be by Tuesday? Right. And then you forward that email to the email that they gave you. And what this is supposed to do is read that email contextually know based on who the sender is, who your customer is based on.

Hector Garcia: Ship me another widget A. It will literally create an invoice and add a widget A and invoice it with a, [00:29:00] you know, shipping date of whatever. Okay. That is where they're heading now. It's active. Now, I haven't tested it. I haven't figured out how this works. We could probably dedicate an episode to that. Alicia, I don't know if you've played with this either. Um, but this is like the first step that I see where actual AI will be used. Because you don't have the historical bank feed data. Uh, of 6 million companies to discern how to categorize [00:29:30] a certain transaction like you have in bank feeds. This is true AI. This is actually reading the email and figuring out with the whatever random words your customer sent you what the instructions are, and that's a really important piece of it. Like now it's going to like it's going to it's going to extract accounting transactions out of the context of an email conversation. So I find that really interesting. And we'll keep you posted when you we actually test this and see and see this work. But this is the sort of the the first [00:30:00] iteration of this. It also looks like the next thing that they're working on is going to be, uh, estimate creation within the same context.

Hector Garcia: So the customer emails you and says, hey, how much do you charge for bookkeeping services? And then you have an item in your QuickBooks file that says, you know, bookkeeping, $500 an hour or whatever it is, $100 an hour or whatever. And then you forward that email to QuickBooks. Maybe you add a little a text that says, hey, um, create an estimate for 17 hours, and then it will take the 17 hours as a quantity, [00:30:30] and then take what the customer asked for, and triangulate that with your item list and create an estimate for you. And of course, give you the choice to check that before you send it to the client, right? That's always that should always be the case. So these are the things that we're going to start that we are going to start seeing. The other thing you see on the on the right hand side, you see this thing that says create new and you can upload, um, a contract, a proposal, something that's a sort of a core document. And it will grab [00:31:00] that document and generate an invoice for you. Okay. This is brand new stuff. Okay. So similar to the email, we'll read the email in context. Now you can read a PDF file or a doc file and convert that to an invoice.

Alicia Katz Pollock: I'm really glad that you pointed that out because I hadn't noticed it before. And one of the things that I've been frustrated by is if I get a bill in my email, I've been able to forward that bill into Qbo and it shows up in the receipts so that I can turn it into a [00:31:30] bill in qbo. But I've been frustrated that I get notices from my customers or from affiliate programs who are paying me out, and I haven't had any way to go ahead and create the sales receipt for that, or an invoice for it, so that when the payment comes in, it's already been registered, and I've had to either use the banking feed or recurring transactions to do that. So I like I want to test this out with some of these especially. It's the beginning of the month, you know, hey, [00:32:00] you're going to get this payout. I want to go here and upload that, that PDF or email that they're sending and see if it will actually create that for me. That'd be really cool. Yeah.

Hector Garcia: And what I find interesting is they started with invoices, not with bills. And you would you would assume immediately that that the, the bills is the first thing they start with. Because every app I mean there's so many apps in the in the ecosystem per se that that's what [00:32:30] they focus on, right? They focus on automating that, you know, the build that they send you or the bill they send you to put it as a build in QuickBooks. But the invoice creation has never really been something that they, that they have automated, which I find that interesting.

Alicia Katz Pollock: What's what else is under that drop down where it says generate new invoice.

Hector Garcia: Yeah. So it says build coming soon. So so that's the other side of the story is you know they're going to do it with that. But there is a version of this already right.

Alicia Katz Pollock: Yeah.

Hector Garcia: But the current version doesn't use AI. So I guess that that's [00:33:00] the difference. You know, the difference is that there is something that kind of reads the vendor name, that kind of reads the, the, the invoice number that. So there's something that kind of does that now, but it truly doesn't like actually use AI to, to, to build it. So I find, I just find that interesting period. Yeah.

Alicia Katz Pollock: No I would love to see even the receipts center get a little smarter too, because right now it's a fairly manual, cumbersome process and [00:33:30] get totally gets the job done. And I'm grateful it's there. So it would be nice if that can get bumped up a little bit too.

Hector Garcia: Yeah. So that's what we see. Now. This is the the latest manifestations of Intuit Assist or or Business Feed. The forwarding your emails to Intuit to see if they can create an invoice for you. The forwarding the estimates to to intuit to see if they can create the estimate for you. There's something about invoice reminders, but I honestly we [00:34:00] already have an invoice reminder system, so I'm not really sure why you would use AI for invoice reminders. We have a complete rule system for that. So I wonder what that's going to be, how that's going to be different. Like I just I just don't see the value proposition on that yet. But this concept of uploading a a PDF or uploading, um, a doc file, it actually says the actual words that it says its contract proposal or similar doc, but it needs to be as as [00:34:30] PDF. So you contract that and I'll discern that and try to create an invoice from that. But as a matter of fact, what I will do is and we'll do this kind of live here, I'm going to upload a, an invoice that I have, and this would be the portion that maybe will do a quick little video in YouTube and show this portion. So this is the invoice that I have. I literally went to Google and typed sample invoice and I just got a random invoice here. Um, that has three lines eggshell something, top cube and color [00:35:00] shirt, floral cotton dress for 500 bucks.

Hector Garcia: I uploaded that PDF into it and then asked, uh, QuickBooks to create an invoice for me and it didn't do anything. Okay, um, let's try this again. Um, and I hate to do this in a podcast because it's just it's just so unfair. But, you know if my first interaction with AI here is that I upload an invoice and it doesn't do anything, then you know who's going to [00:35:30] be excited about that. So let's try this one more time. Maybe it's one of those things where the first time it gives an error. So I uploaded a PDF with an invoice is going through the percentages. It's it's loading I guess it's processing. So this should take um maybe about a minute okay. And there we go. So it went ahead and opened an invoice. But it didn't. But it didn't do anything. Right. So yeah, to be fair, you know, the item list doesn't match what the invoice is. The the vendor [00:36:00] name wasn't even there. So to be fair, there really wasn't any information. But then where is the training aspect? Where's the part where I tell it, hey, this item means this in my item list. This item so I don't I don't even know how this is going to work. So I'm sorry into it. I, you know, we tried it. It didn't do anything. What can we tell you?

Alicia Katz Pollock: But now they have your invoice so that they can start training it.

Hector Garcia: Right? Maybe, maybe. But but the problem is again, this is so the reason why I wanted to do an episode on this. Because there's [00:36:30] I the concept, which is fascinating. There's I that promise in terms of practical, real accounting related solutions. And there's still a big gap. So like to me, all these software companies that are saying we use AI, we add AI, I can do this and that. To me, I'm sorry, including into it. It's just a gimmick. It's just it's just more marketing fluff. And really, the reality is [00:37:00] they just don't want to be seen as being left behind. Oh, zero added jacks or whatever their weird. I thing it is we have to add a weird AI thing to it. So like everybody has to add a weird AI thing that's not functional, not useful to anybody just so they can save face. Kind of like a monkey. See monkey do where? Hey, you know we're not left behind. We are not. You know it mostly. So the stock market doesn't freak out about your company being left behind. So one of the things I really, truly worry about is where is the thought leadership, [00:37:30] you know, on these software companies being truthful about this is what we're working on.

Hector Garcia: It doesn't work well. You were learning together. The net benefit is going to be x, y, z. But their websites don't say that. Their websites say automate this and that. You know, dog whistle fire your accountant I will do this and that. And that's the reality. And that's how that's why accountants are so hesitant to get into AI because they feel it's a big dog whistle to get rid of us and and circumvent us. And [00:38:00] the reality is it can't do none of that. It can pretend to do that. And Intuit is very good at pretending to do things. It can pretend to do that, but it won't really do it. So we need to embrace the fact that it's happening. It's coming. But we also need to be very clear with people that, hey, it's going to get there eventually in a couple of months or a couple of years. But not now. Don't expect it to do that now.

Alicia Katz Pollock: Well, this is where the tail wags the dog. Because on the one hand, if we have people who, um, you know, we [00:38:30] need we need to start learning so that we can help it develop. But it's not going to develop until we help train it for what we need it to do. So it is it is a little circular, but it's definitely an idea whose time has come.

Hector Garcia: Well, we in this podcast will keep you posted of all the developments of AI, relevant and irrelevant, because they're all they're both important. Unfortunately, they're both part of the framing of what AI is and isn't and could be and what they're promising and what's happening [00:39:00] and working and not working and all that stuff. So with that being said, Alicia, what's going on in your world?

Alicia Katz Pollock: Well, I have come back from scaling new heights and then ten days volunteering and playing at Oregon Country Fair and then bridging the gap. And so I've been out of town and conferencing and and just being super active for six weeks. And so what's going on in my world is I'm actually taking a few weeks to just get the work done and lay in the hammock while [00:39:30] I can before it gets cold. How about you, Hector?

Hector Garcia: So yeah. So we're recording this, uh, about two months out from reframe, and that's my conference coming in October in, uh, Hollywood, Florida, close to the Fort Lauderdale airport. And we're very excited. We're, we're, we're selling we're we're going to sell out at some point. And, um, come check it out. Influential conversations for accountants at Reframe 2020 4.com. We'd love to hang out [00:40:00] and learn about how to communicate our value as humans versus I together, I reframe 2024. Reframe 2024. Com thank you everybody. We'll see you on the next one.

Alicia Katz Pollock: See you in the next one.