AI You Can Use Today In Accounting with Giles Pearson
Attention: This is a machine-generated transcript. As such, there may be spelling, grammar, and accuracy errors throughout. Thank you for your understanding!
Blake Oliver: [00:00:05] So instead of creating all these templates for your emails, you can simply create a prompt and then it will customize the email based on the client and the sender.
Giles Pearson: [00:00:15] And John no doubt will use Chatgpt to distill that email down to two bullet points. We're working on your return.
Blake Oliver: [00:00:23] At what point do we have our outlook inboxes replying to each other and talking to each other? Right? Have your have your email client, talk to my email client and set up a meeting.
David Leary: [00:00:35] Coming to you weekly from the OnPay Recording Studio, this is the Cloud Accounting Podcast.
Blake Oliver: [00:00:43] Welcome to the Cloud Accounting Podcast. I'm Blake Oliver. And I'm.
David Leary: [00:00:47] David Leary.
Blake Oliver: [00:00:48] And unfortunately, we have to start this episode with some terrible, just horrible news. Tee Sheets Founder Matt Russell. Is dead in an apparent murder suicide in his home in Idaho. David, you you personally knew Matt Rissell. I met him a handful of times. I didn't really know him all that well. I do know that Tsheets is one of the first real success stories in the whole cloud accounting world. It was one of the first apps to get acquired by Intuit for massive sums of money and. It turns out that the first episode we ever did together of this show, episode three of this podcast, we were talking about t shirts getting acquired by Intuit, and.
David Leary: [00:01:49] It's why you invited me on.
Blake Oliver: [00:01:51] That's why I invited you on the show. Yeah. I also want to welcome Giles Pearson, our guest today. Giles, welcome to the show.
Giles Pearson: [00:01:59] Hi guys. Welcome from from New Zealand and it's great to be here with you.
Blake Oliver: [00:02:04] Apologies that you know, for the I don't know. How do you describe it? Like it's just normally on this show, we don't talk about stuff like this. It's not it's not our usual like we can be naked about things, but it's not we're not usually talking about bad news like this.
David Leary: [00:02:22] And so I wrote a bunch of notes like because I, you know, it's been pretty much dominating my thought process the last 24 hours. And like, if you can appease me a little, Blake, in the listeners, I'd like to kind of like talk about some things. Yeah, Yeah.
Blake Oliver: [00:02:36] So I think it would be great if you could just sort of tell the story of T sheets because it's so tied into. Matt Russell Yeah. Um, first I just want to get the facts out there for the folks who haven't heard, if that's all right. Um, so this comes out of Boise, Devcom, their local paper. Matt Rissell, founder of Eagle Company T Sheets, is dead after a shooting at his Eagle home early this morning. The coroner said he died from a gunshot wound to the head and said it was a suicide. And when the police arrived shortly after midnight, they found Russell dead and another woman in the home badly injured. Who it. I have not seen this in the news, but I understand it was his wife. It was his wife, Robin, who survived but is in critical condition. And. The case appears to be, according to the police, an attempted murder suicide. Officials say it appears Russell shot the woman and then turned the gun on himself. Russell was 46 years old. So with that and.
David Leary: [00:03:39] The three kids.
Blake Oliver: [00:03:40] And he had three kids. Has three kids. Yeah, had three kids. So anyway. David Yeah. Could, would you mind sort of, you know, telling the story of t sheets and why. Yeah. Why this is so important.
David Leary: [00:03:51] Like, first off, obviously, this is like horrible, horrible, horrible nightmare tragedy. Like, no matter how you look at it, like, there's no no way else to look at this. And, you know, you know, there could be an entire podcast or show about guns and mental health and, you know, going down those paths and, you know, but in the end, nobody really knows what's happening happened. Right? And I've gotten texts. There's a lot there's speculation and things like that. And it's like, that's not going to help anybody in this thing to do that pretty much the last day, Right? It's the only thing my wife and I have talked about, right? My wife's met his wife like we've known each other. But I don't know that. And, you know, there's thousands of accountants or accounting and bookkeeping, our community. Right. Thousands of people who have been impacted by this and are have emotions all over the board on this because, you know, everybody has a story or a T sheet story. Yeah, everybody has one. And, you know, I've been really struggling to digest this a little bit the last 24 hours. And, you know, I knew we'd have to talk about it on the podcast, but some part of me was like, We shouldn't even talk about it on the podcast. And then, you know, I've just been digesting this and the thoughts are swirling in my head. The more I thought about it. You can't tell the history of cloud accounting. Like if you're going to write a book, the history of the cloud accounting story, you could not write that book without a chapter about Matt and T sheets.
David Leary: [00:05:07] And that's why it's kind of important to talk about it on the show. Yeah. And I think it's just important to kind of pause and acknowledge the ripple effect Matt and T Sheets had on this entire industry because it's very, very huge. I mean, in fact, you already kind of brought it up. Maybe what we're doing right now, the Cloud Accounting Podcast may not exist, right? If it wasn't for Intuit buying T sheets and you invited me to be on the show. Yeah, right. They may not exist. Right. And acknowledging that and those are the kind of ripple effects you know, I'm talking about. So I'll kind of rewind a little bit because I think you can see where this like other things in the industry and to kind of tie it together. So, you know, I kind of think a lot of these apps I first launched on the QuickBooks app stores, my app babies, right? I personally care about these apps, wanted to see them succeed and they were one of the app babies. And so Matt and I met in like 2012 and I was working with him to get his app launched on the QuickBooks App Store. Right. And it was early 12, 15 apps at that time might be on the App Store, maybe even less, right? It was more than seven. And I at that time Intuit was reviewing every app, so I was like doing a tech review of every single app that existed, right? And probably at that time I had 45 open tech reviews going on and every one of them was a battle, like I'd say, you know, Giles Developer, he has an app.
David Leary: [00:06:32] I would say, Giles, fix this in your app. And the app would like, Oh, I don't want to fix that. And it was just this battle. And I remember I got on a meeting, I tested the T sheets out and how it interacted with QuickBooks, and I got on a phone call with their developer team and, you know, Matt and the rest of the founders there were teeny then, right? And the meeting ends. They send me an email with their plants 3 or 4 days, they fix it all. And I instantly recognized that they were different from all the other companies I was working with. You know, T was like special. The company themselves were special from day one. And so not long, much longer after that, somehow or another, I don't even know how or why. I was in San Francisco. They were in San Francisco. We were all at the same hotel and it was not an event. It was like early days. There was no QuickBooks Connect. There was none of these things that existed. And basically the So it was Matt, it was Brandon, the other founder, and Jennifer who led the marketing team there. It was the four of us just sitting around at this hotel bar table.
David Leary: [00:07:25] And basically we drafted up the playbook on how to win the cloud accounting app space, right? What conferences to attend, who to connect with, who to partner with, what swag to give away, why you have to win with these accountants and bookkeepers and ProAdvisor, right. Why they're important. And then the rest kind of is history from there, right? I think at that point now the rest of the world starts discovering and touching with t sheets. But think about it, right? You like that shirt you're wearing? Those from whatever app that's from t sheets started that these these cloud accounting podcast. This is just me ripping off t sheets right at some level right. That gift bag of swag some apps sent to your house that's T sheets the number one accounting bookkeeping podcast in the world. Dude that was a total Matt move. Matt Like put a I just copied t sheets when they said they were the number one app. Right? Like these are this is big. All those over-the-top conference parties. It's t sheets having your app, your sales and service team trained and certified in QuickBooks or zero T sheets started doing that, first showing up to Tucson to the Intuit Call center and giving out $100 bills to two employees. Right. That save T sheets on a phone call or your app name on a phone call. That was Matt first, right? He was doing that before anybody else. All you app developers that are sponsoring for example, Joe Woodard scaling new heights. Often you hate how you can't have a gathering or a party or an event within 50 miles of the venue.
David Leary: [00:08:42] Right. Within ten days of the event, that's the t shirts rule. It's because Matt and T sheets threw these crazy parties. Getting reviews on the QuickBooks App Store. They're by far the first app to hit 1000 reviews. The miles ahead of everybody else. And they they set the bar for that. I'm probably forgetting like ten other things. Another thing is like gusto. Zenpayroll and gusto may not exist where it's at today. If it wasn't for Tsheets. They're basically Long story short, there was a bug in QuickBooks that didn't let developers like Tsheets who are tracking overtime send the overtime data to QuickBooks for payroll. And so Tsheets had all this momentum and all these customers. And so they would just. Well, our integration with Gusto works and they would just send them at that time. It was then payroll. And so the first initial waves of customers that were going to gusto and payroll were coming from Tsheets. So apps you're using, right? So like I think like, I don't want this to become, you know, the other thing to think about too is like all these apps you see doing things right, they're, they're just copying Matt and Tsheets. They might not even these new apps that are on the market today, they might even know Tsheets even existed or witnessed Tsheets because they're just copying the third or fourth variation of somebody who copied what Tsheets did in this space, right?
Blake Oliver: [00:10:03] They yeah, they created the playbook for how to build. An app that solves a particular problem in our profession that integrates with QuickBooks or with Xero now.
David Leary: [00:10:14] And I don't want this to get confused with like praise and because people are questioning, you know, murder, suicide, how do you praise somebody? But I just want everybody to pause and acknowledge that something you're using today, something at your desk. Right. Something you see at the next conference is probably a little piece of Matt in it. Yeah. It's just like that's the way to think about this.
Blake Oliver: [00:10:32] And the reason this is hard, like, I didn't know him that well, but he was incredibly charismatic. And even just meeting him a few times, like I could feel, you know, positive energy. And that's what makes this so hard is I can't reconcile the person I met with what happened. And I think that's what a lot of folks are going through. Yeah. Um, just to put a number on what you're talking about, David, In 2017, Tsheets was acquired by Intuit for $340 Million in Cash and Other Consideration. It was the first really big acquisition, the first really big, you know, financial initial impact. And probably that in the mind of investors is what stimulated the whole economy that we have now in the accounting technology space.
David Leary: [00:11:18] Yeah. And I think it also like the it proved that because they were in Boise, Idaho for the longest time, if you were going to be a tech company, you have to be in San Francisco. And it kind of proved these apps were just showing up in all these other towns and these tech hubs started appearing. Yeah.
Blake Oliver: [00:11:35] So. I know it's going to be weird, David, to, like, change your mindset and shift and just go now to our regular show.
David Leary: [00:11:42] Like, it feels good to be able to communicate that out because it just it's. Yeah.
Blake Oliver: [00:11:47] So, you know, you know, it's hard. You're grieving. You knew him well. But we're going to we're going to change our mindset. We're going to pivot now and we're going to we're going to do our regular show and yep, yep, yep, yep, yep. Okay. So I don't know, let's let's talk about something positive, right? Something that's actually very exciting and positive and. Gets me. Really excited. And that's Chatgpt. And I and I know everybody's been talking about this a lot. But we've got some practical applications of it. We've got demos of it. We've got it in the practice management software now, and I've brought some videos. We can actually watch it in action. So if you haven't seen it or tried it for yourself. And you're watching our live stream. You're going to actually see it. And if you're listening to the podcast episode, you could go find it on YouTube and watch that as well. We'll do our best to explain what's going on. And Giles, I think this will be interesting to you because you have been doing your own experiments in AI, so you run Accountests, which is a well, how would you describe Accountests, Giles?
Giles Pearson: [00:12:51] Uh, so we, we help public accounting firms avoid bad hires. That's that's our raison d'etre, I guess. And we do that using skills tests and personality profiles to help you make sure you understand the candidate that you're looking to bring on board.
Blake Oliver: [00:13:11] And you've been playing around with AI by feeding it questions from your not your personality test, but your skills test, right?
Giles Pearson: [00:13:19] Yeah, that's right. Look, we've we've been playing with it to try and help us develop some new questions and that's actually hasn't worked that well, to be fair. But we've certainly been feeding it questions both from our CPA level tests and our sort of bottom end debit and credit test and just just seeing how how it goes and whether it, you know, where its strengths and weaknesses lie. And I guess, you know, it's just fun at this stage. But actually the fun is is getting serious and it's getting serious pretty fast.
Blake Oliver: [00:13:53] How is it doing? Like. Like if if you had to rank chatgpt as a higher how is it doing against a typical applicant for a staff accountant job, for instance?
Giles Pearson: [00:14:02] Well, so so we had it set our CPA public accounting test. So that's a bit of tax, a bit of advisory, a bit of reporting, a bit of sort of core accounting stuff. So, you know, what you'd expect, say a client accounting manager or to to be all over in most public accounting firms. And it was on the 93rd percentile. So yeah, in the top, top 7% of the norm groups. So behind all of these tests is is a norm group of a whole bunch of people who've who've set these tests in the past. So yeah, it would you would certainly choose it as your well likely choose it as your preferred candidate in terms of its technical understanding. And Charles.
David Leary: [00:14:44] Did you do any training of this or did you just upload your questions and ask it for answers?
Giles Pearson: [00:14:49] I just uploaded the questions and asked for the answers.
David Leary: [00:14:51] So this was just already in its basic train module training that it's had.
Giles Pearson: [00:14:56] But but it learns. So one of the questions we put up it it got it wrong and I went back and go, This is not right. And this is why. And it goes, Oh, you know, look, I'm really sorry about that. You know, I'm trying to get better all the time. And, you know, so I said to it, Well, so will you get it right next time when I ask? And and. I then got one of my team members in the Philippines to upload the same question a week later and it got all right. So. So it's learning as well.
Blake Oliver: [00:15:27] It's learning. So it's it's amazing. And I've been talking with folks all week about AI in audit, in tax, in mainly in audit and tax, also also A. And I am just amazed at what it can do and I'm really excited. We're on version four, GPT four like it's been like four months and people are already finding practical applications for it. So.
Giles Pearson: [00:15:52] Oh, look. And there's and there's some really exciting stuff coming down the pipe. Right? So certainly here in New Zealand, I know there is a product, you know, look, it's not imminent, but it's but it's, but it's close, you know, and what this is going to be able to do is actually do that initial review of your client. So at the moment, you know, you get the you get the the file in the door and somebody has to go through it and work out where all the bones are buried and basically go through the checklist. You know, everybody's got a checklist. So they basically automating this checklist using AI. And so it will be able to. Sorry, David.
David Leary: [00:16:36] I was gonna say so like the, like the file review the first time you sometimes before you even take on a client, you're like, let me just look at your file and kind of analyze it so I can give you a price. Well, I think this is.
Giles Pearson: [00:16:46] This is more than that. This is this is the this is the initial review, say, of a year end file. So it will be able to go through compared to last year, drill into each of the accounts. So, you know, say repairs and maintenance or subscriptions or whatever you particularly want to drill into. If the invoices are uploaded, go and go and interrogate the invoice, propose journal entries to fix errors that it sees and be able to flag transactions that might have a tax impact so that you'd be able to go, Here's an adjusting entry on my tax return. So essentially. You know, you take that to its logical conclusion. The first time a human needs to touch that file is when the manager gets the file review with the proposed journal entries and hears the questions where you need to dig deeper or you need to think he's a he's a transaction that looks like it might be capital, but they've the clients coded it to to an expense. You know, here's my take on what it should be. Here's a copy of the the invoice. You make the decision, you know, click yes or no balance sheet that you know. Yeah. So so so that's that's that's not that far away.
Blake Oliver: [00:18:10] And we should say that, Giles, you are running a contest, a technology company helping accountants hire better, but faster and with confidence, I think is is your tagline. But. But you have a long career in accounting. You were a partner at PWC in New Zealand.
Giles Pearson: [00:18:30] That's right. So I was a partner for 18 years and with them for 25 odd years in an office, you know, about the size of your local shoe store. So. So we had 45 people in our office, sort of four partners. It was like a mid-sized firm, I guess, in regional USA. So, so the experience, I guess, of the of our working environment was very similar to to many, many accounting firms across the world, but overlaid that was the PwC experience, I guess, which comes with its ups and its downs. But yeah, you certainly learn a lot and you and you and you're working in an extremely supportive environment.
Blake Oliver: [00:19:13] So, so. Pwp A little bit different in New Zealand perhaps than in New York City, right? But you know, still Big Four and it's relevant because we just had a question come in from one of our live stream viewers. Do you think the shortage is offset by outsourcing? And I talking about the talent shortage in accounting? Big Four pushing layoffs has me nervous. What do you think about that?
Giles Pearson: [00:19:35] Giles Well, you know, I think AI is going to be challenging for the Big Four, but for all the profession, because what AI is going to do is deal to all of those jobs that the interns and the college graduates do for their first 3 or 4 years. And, you know, you've you've done audit. Blake does.
Blake Oliver: [00:19:59] One audit. Okay, that was enough.
Giles Pearson: [00:20:05] Yeah, I hope it was a big one. Um, you know, AI is going to going to chew all of that stuff up real fast and you know and so I think that's, that's a positive in terms of the talent shortage, although I think we're probably there's a time gap there. We're probably 3 to 4 years before that becomes a more serious issue. But I guess my biggest concern is how do we train our people who need It's not that they need to do three years of mundane work. You know, three months of mundane work is probably enough to ground you into some of those things. But we do need them to get experience and we need them to get that experience in a safe environment where actually they're they're not out there giving their opinions to clients when they clearly don't have the capability of doing that. So how do we how do we bring people through? And I think that's a that's a big challenge for the profession. And I worry whether our educational institutions are up for that challenge.
Blake Oliver: [00:21:07] I say that my entire career has been built on Google searches, knowing what questions to ask, and I don't see accounting programs. Now. Look, there could be great ones out there. There could be not so great ones out there, but I think on the whole, most accounting programs are not teaching, not teaching their students to ask questions. Right. It's here's here's what I'm going to teach you. You're going to repeat it back to me in a lot of cases. And we should be questioning accounting standards. We should be questioning tax laws like why do these exist? And whenever I did that in my classes, my professor got really annoyed because they didn't really want to have to explain or maybe they didn't even know the answer or had never thought of it. Right. Maybe it's frustrating, right? When stuff doesn't make sense because a lot of accounting and a lot of tax doesn't make sense.
Giles Pearson: [00:21:59] And look, I think communication is so so asking the questions is is a big part of that. But it's about that ability to communicate. And it's so rote learning is, you know, is gone really, I think as a way to to to help our people be able to work in the modern world and what they need to be able to do. They need to know stuff because you can't you can't communicate with the client if you don't know anything. But it's about combining those two things.
Blake Oliver: [00:22:32] Yeah, it's not so much about, you know, memorizing all the details because you can look them up whenever you need them. It's more about understanding the concept.
Giles Pearson: [00:22:40] That's you need to know some things, you know. And I think that's and that's where we would come from, I guess, as a as a testing company, if you want to hire a candidate, you still need a candidate who actually knows some stuff. But you know, that's interlaid with their ability to then communicate. Because if you're in a meeting with a client, you know, you can't you need, as you say, you need to know the concepts and you need to know enough to have a discussion. Where can I.
Blake Oliver: [00:23:07] Have chatgpt open on my laptop and be asking it questions.
Giles Pearson: [00:23:12] If if you don't want to be paid? Yeah, sure. Yeah, but.
David Leary: [00:23:15] But. But. Even if you do that right, you still have to have you somewhere along the line, you have to get the knowledge and the experience because it's easy. Chatgpt will make a mistake and then you'll just repeat it out loud. Then you'll make a fool, right? But again, it's kind of like maybe right now you ask your intern to research something, and then if you don't have time to validate it, you might say that what your intern did. So it's just I think your intern might come back. Not as confident with the information in Chatgpt always comes back. Pretty confident that.
Blake Oliver: [00:23:43] Very confident you got you got to take it with a grain of salt. So, hey, all this discussion of chatgpt has stimulated some conversation from our livestream viewers. Reminder to our podcast listeners that you can also watch us on YouTube. And if you subscribe to our channel, you can get notified when we go live and see what we actually look like. Sometimes people are surprised.
David Leary: [00:24:02] We think bigwigs in the chat this week, well.
Blake Oliver: [00:24:06] We have the famous Greg Kite CPE, he says. I asked Chatgpt to write the lyrics to a song and include the words Vacation, Avalanche Strength, Love and Trust. Not only did it include three verses, a chorus, a bridge and an outro, it was so good I literally got choked up. I performing creative tasks at that level makes me realize that we will all be unemployed by the end of the year.
David Leary: [00:24:30] Well, yeah, I'm going to use AI to write to tell it that I'm the an accounting comedian and I'm going to have it write me some jokes.
Blake Oliver: [00:24:39] Well, I tried having chatgpt write me a joke for my next presentation and it was probably the worst joke that I've ever heard. So I think Greg Kite's job as a comedian, CPE is very safe, I would say, based on that experience. Chris Maxey said. Boston Dynamics added Chatgpt to their robot dog, so it won't be long before the Big Four hires those robots to go into the office, since they won't complain about that. Or they'll also get burned out and will result in the Black Mirror episode. I have not yet seen Black Mirror, but I can only imagine what a sentient or not not necessarily sentient, but autonomous Boston Dynamics robot dog could do. Yes, that will be a frightening future. And Christopher Perez said, soon comes the day when sentient AI starts quiet quitting. Jennifer Johnson said, I had my class use it. Jennifer is a professor at the University of Texas. The better their questions to it, the better their paper was. Some loved it and some were really frustrated. Great learning experience. Plan to keep using it. Awesome. Well, I'm glad you haven't banned your class from using it. Jennifer. I think that the school districts that did that made a big mistake. That's like banning calculators in 1986. Remember? That was. Yes. David, you brought that story to the 82.
David Leary: [00:25:55] It was Yes. Yes.
Blake Oliver: [00:25:56] Oh, 82. Yeah, something like that. Right. And Michael Allen says, I can be astounding but can't ever be astounded. Oh, I'm going to have to think on that. That's a good one. I might have to steal that, Michael. I'll be sure to give you credit. All right. So shall we see some AI in action? Because that's what's exciting about this week or the past couple of weeks is we've actually had apps integrate GPT into their software. Yeah.
David Leary: [00:26:21] When you start to bring those up, like I'll tell, I use the Chatgpt this week on my own to let to take unstructured emails of of appointments. I like to have flights on my calendar if I'm going to an event. I like the start and stop times of all the the agenda on my calendar and it takes a little bit of time. You're like, you have it up, you create the event, create the next event, right? And so I was able to take some emails that don't have, you know, where the address is in one paragraph of the location or two different addresses. And then there's an agenda or my flight information from plane tickets, hotel reservations. And I basically interacted with Chatgpt to where it was able to give me the AI, the Internet, the Excel file. Well, it doesn't give you the file. It just gives you what to copy. It gives you the.
Blake Oliver: [00:27:05] Code, gave you the code for an Excel file.
David Leary: [00:27:08] And then I just import it in my Google calendar and they're all there. And that's pretty cool. And it was and I and I used the intern interaction because it's kind of like the intern. Let's say they put it in your calendar, then you're like, Well, you know, on airplane flights, I like to have the airplane emoji like the departure city, the airplane, then the landing.
Blake Oliver: [00:27:23] In your calendar.
David Leary: [00:27:24] In my calendar, yes. And I told it. You're demanded it and then it did it and put it in the iCal file. But then you can take it a step further so I can just paste in the flight info from Southwest in there. Right. But then I'm like, I also when you do that, I need put in an appointment for an hour, two hours before the flight to drive to the airport. So it's putting that appointment in for me, right? It's putting in oh, check in for the flight the day before. So instead of me put one airplane flight results in three calendar appointments, instead of me doing all three of those, it's able to do that. Right? And so it's a lot of back and forth. But if you could just give you the file and the next step, I guess, would be how do I connect through Zapier? So it just puts it in my calendar.
Blake Oliver: [00:28:04] And that will be possible very soon with the plug ins feature, which I'm on the waitlist for, but I still haven't gotten access to it. But apparently you can connect Chatgpt to Zapier and then if you can connect something to Zapier, basically you could do that. David You could have it create the event on your calendar.
David Leary: [00:28:20] And at first there was a little bit of dancing. At first it wouldn't import the file in. It wasn't correct, it wouldn't import it in. So I used like some website that was like a iCal validator and it gave me 28 errors that they that that chatgpt did wrong on the file. So I took all the errors and pasted them back in and it fixed them all and then the next five took and then I imported it in. It was pretty impressive. That's amazing. Yeah.
Blake Oliver: [00:28:43] All right. So let's check out these four different videos. They're very short, and I just cut it right to the part where we're seeing the AI in action. So the first one is going to be Client Hub because I think they actually might have been first to this. And what they have done is they have created a it's like an you ask the AI to create you a checklist and then it goes and creates you a checklist. So this is Judie McCarthy, the founder of Client Hub.
Judie McCarthy: [00:29:10] So previously we had a pretty good library of job or workflow templates here that you could select from. But now we have unlimited templates because we have magic workflow. Magic Workflow uses AI power to allow you to create a job just based on a description. Let's give it a try. Let's say I'm going to create a job to post sales from Shopify into QuickBooks. Okay, let's push the button and see what comes up with.
Blake Oliver: [00:29:38] And all she did was she said, push Shopify data into QuickBooks. She wrote that into the job description.
Judie McCarthy: [00:29:45] This could take up to 30 seconds for to create that job for us automatically. But when we get there, you're going to see look at this. We have the job. We have the tasks. And when we click on those tasks, we even have the task details. So we're going to export sales data from Shopify to a CSV, import that into QuickBooks using the import data feature, match the sales transactions and QuickBooks with the corresponding bank deposits. This is amazing.
Blake Oliver: [00:30:18] And then create sales receipts, record fees and expenses and reconcile accounts. So based on that initial description, just post sales from Shopify into QuickBooks. Gpt has created 6 or 7 tasks with descriptions in those tasks for more detail. And that's a great starting point. And I love this use case because it's one of the reasons that a lot of firms get stuck implementing practice management software is they realize as they're implementing that they don't actually have any processes and then they get stuck creating the processes while they're trying to implement the software and then nobody ever uses it.
David Leary: [00:30:55] And the magic here. Oh, good, Joss.
Giles Pearson: [00:30:58] I was just going to say, the staff get frustrated and, you know, you've always got those those people in the team who are slow adopters and they drag the whole team down and yeah, the whole implementation either just, just drags on or, as you say, people just give up.
Blake Oliver: [00:31:16] So the next one.
David Leary: [00:31:17] Or before you jump to the next one, like I was saying. So the beauty of this is like having this built in the app is the real time savings because I just opened up what you're doing the video, and I went to Chatgpt and I said, Post sales from Shopify into QuickBooks. I have the same kind of list of steps, but now I got to take them and manually copy paste one line at a time into the other app, which is just bananas crazy. But now you could just use the app and do it and I don't even need to have a chatgpt at all. Right. And that that's where this is going to be interesting is people are just going to be using it like that in many apps, right?
Blake Oliver: [00:31:54] So that was creating tasks in your task management, workflow management, software and Client Hub. Now let's look at what Canopy did with helping you draft emails to clients.
Canopy Clip: [00:32:06] Communication today is so vital. If we don't communicate clearly and efficiently in the beginning, we can find ourselves wasting time afterwards with answering, clarifying questions, redacting statements, not having fully thought through the communication that we needed to deliver in the first place.
Blake Oliver: [00:32:26] And I'm going to skip ahead a little bit in Blake.
David Leary: [00:32:29] Yeah. Before you skip ahead, you pause is so on the screen right now. Is this a create email screen within Canopy or is this like just an email?
Blake Oliver: [00:32:38] This is within Canopy.
David Leary: [00:32:40] This is within Canopy. Okay.
Blake Oliver: [00:32:41] Canopy like Karbon allows you to use it as your inbox, which is, which is great because often you're creating tasks, you're creating work, you're referencing work from your emails, so it makes sense to have it integrated into the tool. Got it. But of course the challenge with that is you don't get all the benefits of like plug ins for Chrome or for Outlook or whatever. So this is why they're having to build this. So let's watch it. Help to create a draft.
Canopy Clip: [00:33:07] An email Canopy has a new feature that will allow you to draft an email based on the bullet points that are in your brain as to the things that you want to cover in a conversation. For example, it may be that more work was required out of scope, higher price. You okay?
Blake Oliver: [00:33:32] So that's all she wrote? More work required out of scope, higher price. You okay and then is pressing submit.
Canopy Clip: [00:33:39] Put this into the generator the prompt box and it will draft an email for you.
David Leary: [00:33:45] So it's showing it working. There's the.
Canopy Clip: [00:33:47] Hey Brad, I hope you're doing well. I just wanted to check in with you regarding the additional work we discussed for your project. It outlines the whole thing for you in a very professional way.
Blake Oliver: [00:33:58] So going on reading this email, right, I wanted to make sure that you're comfortable with the new price and that we're still within your budget for the project. Please take some time to review the new quote and let me know if you have any questions or concerns. We're always happy to work with you to find a solution that works best for you. Thank you for your understanding as we look forward to continuing to work with you. Best regards, Casey. So it wrote the whole email from just the bullet points, just the comma separated bullet points. And we know that, like you said, David, you could just go into Chatgpt and you could do that, but then you have to copy and paste back into your email, which kind of eliminates the time savings. So I thought this was so cool. And then Carbon came along this week and took it to another level. So let's watch what Karbon has done.
Karbon Clip: [00:34:44] All right, so here I am inside triage.
Blake Oliver: [00:34:46] And triage is Carbon's email inbox.
Karbon Clip: [00:34:50] First thing I hope you'll notice if you scan down the right hand side is the priority labels. So you can see that I've got a couple of emails that have come through and automatically been tagged as high priority or low priority.
Blake Oliver: [00:35:00] And the way they're doing that is with sentiment analysis. So they look at the body of the email and determine whether or not somebody is upset or if something is urgent. And this is genius.
David Leary: [00:35:11] Clients are going to figure this out and they're going to they're going to use Chatgpt. What's the best way for my accountant to read my email? And it's going to say, Use all caps and use these three words and then you're always going to be able to be the most urgent.
Karbon Clip: [00:35:24] We'll be introducing also a filter list at the top so you'll be able to quickly filter and see what all of your high priority messages are. Now, if I want to action one of these emails, I can open it up. You can see here I've got an email from my client, David Buck, and he's telling me that there's some issues with his billing and I have the option here. Let me just move this around so you can see there's quite a few messages. Just to summarize by clicking the summarize button, it provides me again with a bulleted list outlining all of the content that's occurred in the email message.
David Leary: [00:35:54] So can you like summarize the bullets there?
Blake Oliver: [00:35:56] Yeah. So Client wrote the email and it's multiple paragraphs and we've all had this, we've all had clients or colleagues write as ridiculously long emails and you open it up and your eyes glaze over and then you close it and you don't open it and you open it up later and you think about reading it and then you don't you close it. So what happens here is that the user press, the summarize button and we got four bullet points and the bullet points say David noticed a discrepancy in his billing statement and is concerned about it. David is a valued client of Annapurna and appreciates their professionalism and expertise. David requests prompt attention and a clear explanation of how the discrepancy occurred. Sarah asks Simon Hilton for assistance with the urgent matter. Sarah is a team member, so it summarizes not only what the client said, but also what the team has done in the thread.
David Leary: [00:36:44] This is I'm just thinking like, why is not every email client doing this already? Like every email client should just summarize every email into three bullets like this.
Blake Oliver: [00:36:53] Well, I'm willing to bet that outlook will be doing this real soon. Yeah. As I've talked about on the show before, Google Chat is already doing this with chat threads. So if I walk away from my computer for a few hours and then my team. Is going back and forth, back and forth, back and forth on something. When I come back, I get a nice little summary of what they talked about at the top of that thread. So same idea. Let's continue.
Karbon Clip: [00:37:16] You can do that on all emails, not just the ones with high priority labels. So I'll give you another example of that. We'll grab one here. It's another email from another. David Buck My test data. David's been very busy sending me emails, but again here you can get the other summarized bulleted list. The summary does include comments. It's not just the email messages from your clients, the emails between your team members. It will include all of the comments and actions that have occurred within the email conversation. All right. The next use case, the feature I'd like to show you is the email task. So I'm clicking over here to a work item and I'm going to go over to the task list and you can see we've got this new email, this new task type, which is called email task.
Blake Oliver: [00:37:55] And the task says update client on timeline. So this is part of A. They don't call it this in Karbon, but it's like a project and it's a task in the project and the task is update client on timeline. So you know, you've started a tax return and now you need to update them.
Karbon Clip: [00:38:11] I open it up and it's got a type, so this is the prompt type and you can see there's some prompts in here. So this task is really about letting the client know that the work has started.
Blake Oliver: [00:38:20] And the prompt says, tell the client A team has started working on their tax return. We expect it to be completed within one week.
Karbon Clip: [00:38:29] So I go and I create draft. It's off creating the draft for me. It puts it on the timeline and you'll see here it's suggesting the client that I should send the email to John Smith and it's produced all of the text for me.
Blake Oliver: [00:38:42] And it produced an email that says, Dear John, I hope this email finds you well. I'm writing to inform you that our team has started working on your tax return. We are committed to providing you with the best possible service and ensuring that your tax return is accurate and complete. We anticipate that the tax return will be completed within a week and we will keep you updated on the progress. If you have any questions or concerns, please do not hesitate to contact me. Thank you for your continued trust in us. Best regards. Sarah Gopal. So instead of creating all these templates for your emails, you can simply create a prompt and then it will customize the email based on the client and the sender.
Giles Pearson: [00:39:18] And John no doubt will use Chatgpt to distill that email down to two bullet points. We're working on your return.
Blake Oliver: [00:39:25] At what point do we have our outlook inboxes replying to each other and talking to each other? Right? Have your have your email client, talk to my email client and set up a meeting.
David Leary: [00:39:36] I mean, it's kind of happening already, right? Like people are using all these automated tools to do drip campaigns and those types of things. And then on the other side, everybody's using automated tools to filter spam bucket those emails like it's already kind of happening on both sides, right? This is just the next extreme of it.
Blake Oliver: [00:39:53] Christopher in the live stream says, Amazing. If you can tell it to be direct or deliver softly to Dr. Towards Love languages and actually yes you can. I didn't show that part of the canopy demo but you can tell it to revise an email if you think it's not appropriate and you can say be more direct or be more friendly or be more professional and it will do that. So let's continue with this Karbon demo because I think there's a little bit more.
Karbon Clip: [00:40:19] Again, I can edit this text if I want to change anything, or I can change the tone by using the tone change. Here we go.
Blake Oliver: [00:40:25] Tone changing feature.
Karbon Clip: [00:40:27] So this is just the beginning. We expect to introduce more features and ways to improve efficiency and workflow within Karbon using AI.
Blake Oliver: [00:40:36] All right, so that's Karbon, by the way, If you don't have Karbon, if you don't have Client Hub, there's two options for you. If you don't have canopy as well, there's there's two options for you that work with. I know they work with Gmail. I'm not sure if they work with Outlook, but they're web based. So if you have outlook on the web, you could use them. Grammarly has a plugin for Chrome and you can compose emails with it. I'm doing that already. It can do basically the same functionality where you can give it a list of points and it will create the email. It will also read your emails and suggest responses and you can then change the tone. There's also Jasper. Jasper has a plugin for Chrome that allows you to. Basically have it do anything it can write anything in any text input field. And so it can it can do that. And it can also rephrase what you've written. And I use it all the time. I'll write an email and then say like, make this better and then it generally improves it.
David Leary: [00:41:30] Giles are using. I know like I talked about, I'm using an I to do some stuff. Blake. Blake uses it now on a regular basis to write emails and things like are you using AI, like outside of you playing with it and testing it like you on a personal level? Are you using it?
Giles Pearson: [00:41:43] Yeah, if I'm composing, I guess I don't do a lot of sort of regular emails if you like, like the ones we just looked at. But so if I'm composing, I actually still will do that. I guess that's just the way I think I type and think at the same time. But when I've finished that, I then put that into into Chatgpt and as Blake said, get it to improve the tone and, and you know, make it, make it more, make it more readable. So improving the output. But in terms of developing the content, I guess that's that is still a bit that yeah, it's the same as that communication piece is the bit that us humans still have to have to work out.
Blake Oliver: [00:42:26] I've got one more real quick financial sense. It's financial sense is the last one I wanted to mention. They also have their practice management solution. They've also integrated with GPT similar to Client Hub. So if you're on our live stream, you can see this. They've got this create a template using AI feature where you type in something like create a checklist for monthly bookkeeping for a construction company and then you hit generate and it creates all the tasks based on that. So reconcile bank and credit card accounts, it'll say. Let me pause this. You can't pause an animated GIF track and know all the job related expenses. Anyway, it's a customized checklist with like 12 things for a construction company that would be customized to that.
David Leary: [00:43:13] I dropped in the private chat. Blake. A link.
Blake Oliver: [00:43:15] Private chat?
David Leary: [00:43:16] Yeah.
Blake Oliver: [00:43:18] I'll pull this up. So.
David Leary: [00:43:19] So this is at some level, like, this is a genius post from Expensify at some level. So expensify everybody vanity searches, right? So they they expensify wrote a whole blog post about how they asked Chatgpt what the best expense management app is. And of course, Expensify came back, number one. And what they did is they did variations of it asking this over and over again. So basically they have a screenshot of chatgpt, but then they would type in a summary of what Chatgpt said. So now what they basically did is they wrote a whole new blog post about how great Expensify is. So when the next even in their conclusion, they say it's very obvious this is being trained by what's on the web already. And now. Now they have a whole nother blog post that talks about why it's the best product like they. So the next time the model is retrained, it's going to see this blog post and just reinforce its learning. I think it's kind of a genius. Blog post by expensify here.
Blake Oliver: [00:44:19] So we'll we'll I manipulation become the new SEO right creating posts so that you know like bing bing now searches the web and integrates with chat GPT with GPT four so you could you could manipulate the results by creating a post specific to a certain question. I guess it's the same thing. And that's what they're doing.
David Leary: [00:44:41] They're saying Expensify dominates the expense reporting landscape in English speaking countries. Expensify partners are the most influential people Like it really is. Like this is really just a genius blog post. They gave them a way to write about themselves without writing about themselves. It's really smart blog post.
Blake Oliver: [00:45:00] I've got one more app for you. David. This is not practice management, though. This is. It's an app called Sturppy. Have you heard of Sturppy?
David Leary: [00:45:10] Hopefully I've never caught that.
Blake Oliver: [00:45:13] R.p. It's not a disease. It's not a pharmaceutical. It's financial modeling for startups. Use our proven templates to build a model fast without any prior experience or expertise in finance. Share a live link with potential investors or export the entire model to a formatted spreadsheet with a single click. So we've seen a ton of these startups. And the dashboard is a financial dashboard, financial modeling, financial modeling, replacing Excel for doing financial models. And what caught my attention about Sturppy is that they now have what they're calling their Gpt4 virtual CFO. So I think the best thing to do is just simply to see it in action.
John Ladaga: [00:45:54] Hello, founders John Ladaga here, co-founder of Sturppy. Today I'm super excited to announce a brand new feature just coming into Alpha that we've released for Sturppy Plus. So we are releasing today CFO chats. Cfo Chat is basically a personal CFO available 24 over seven. It's trained on your live data, so you connect your Xero, your QuickBooks account, all your accounting data, your analytics data, and then you can actually query that data, provide insights, create graphs, everything that you potentially would want from a data analyst or a fractional CFO. Our platform is kind of seeking to solve those use cases. So let's start with some basics. Let's just kind of ask it to surface some KPIs. So let's ask it, you know, what was our total revenue last month? Awesome. So our total revenue last month, today is April was was $91,000. We can also ask if things in the past so we could say, you know what was our total revenue four months ago? Total revenue four months ago. December 20th, 22 was 25. K So that's just surfacing some KPIs, but we can actually use CFO chat to actually create some graphs. So let's say create a graph of our total revenue and total expenses.
Blake Oliver: [00:47:17] And this is the part where I started to pay attention because it produced a nice graph, a line chart. Month by month with the total revenue and total expenses.
David Leary: [00:47:28] Just right there and actually has an image like you think that's the important part here.
Blake Oliver: [00:47:32] It's an image, right? So no Excel graphing required, no exporting of the data, nothing of that just instantaneous chart.
John Ladaga: [00:47:39] Great. We've got a graph here of our total revenue and total expenses with the data that we provided it. And it's looking accurate here. So if we remember, we asked it, you know, what was the March revenue 91 K we're seeing that we got March revenue 91 K and these graphs, you can add these graphs over to your live dashboard views and thus be able to view it there. We're going to add the ability to kind of send this data to Google Sheets or Google slides if you want to include this. But let's say maybe you wanted this in a bar chart and don't want to look at it as a line chart. So we could say, you know, change that graph to a bar chart.
Blake Oliver: [00:48:17] So no need to understand how their software works. It just does it.
John Ladaga: [00:48:21] And great, just like that. Now we've got a bar chart of our total revenue, total expense. Could you pause it, Blake?
David Leary: [00:48:26] Did I catch it? Say this correctly? Once he had it done, you could just drag it out of there to your dashboard. So. So you're interacting with plain text to create the perfect chart or the perfect graph, and then you just drag it to the dashboard, right? Wow.
John Ladaga: [00:48:40] Let's change the format of it. So let's ask it something a little bit more complex. Let's do some what if analysis here. So let's pretend we're going to try to potentially ask our CFO chats about a new hire that we're planning to make. So let's say we're thinking thinking about hiring a new operations manager at 5KA month starting next month. Can we afford it now?
Blake Oliver: [00:49:10] This is a question that's a challenge, some analysis here.
John Ladaga: [00:49:14] So it's saying to determine if we can afford hiring a new ops manager. Let's take a look at our current financial situation. So our bank balance in April is 23. K if we add the new operations manager to our our current expenses, we can estimate our future expenses. So the total expenses for April currently are only $1,000, Adding the $5,000 salary estimates that our expenses would be 6900. Considering our current bank balance and the estimated expenses for next month, we can afford to hire a new operations manager per month. But it's also giving us a caveat. So it's saying however now.
Blake Oliver: [00:49:48] So this is where I started to see the limitation of this, right? It's not quite there yet because all it's doing is taking last month's expenses, adding in the salary of the new hire and then saying, oh, the bank balance is $23,000, right? We can afford this. It's not actually modeling out anything at this point. Right. But I mean, you could see how it could get there, I think.
Giles Pearson: [00:50:11] Well, when you think that Chatgpt can work out, for example, and these are these are in our tests. So we we we got it to work out the working capital. We got it to work out, you know, receivable days. So when you can start getting this sort of model to do that for you, I think then it becomes, you know, you really are providing that sort of virtual CFO service, you know, at a much higher level. Yeah.
Blake Oliver: [00:50:44] I think that they'll get this to the point where in Excel you'll have a plug in that will be able to create and update financial models just like an Excel wizard does now, like a person who's very well trained on it. I wonder if that like job of Excel jockey is going to go away. A lot of it will be completely automated. These people who do financial models, analysts, right. Financial analysts, that's been like one of the fastest growing jobs over the last ever since the computerized spreadsheet was invented. Right. We got millions of them now or a million, I think. You don't have to.
David Leary: [00:51:21] Memorize the formulas anymore. Right? You know, if.
Blake Oliver: [00:51:23] You want you don't have to know the formula. You just say what you want. So you still have to know what you want, right? You have to know that this is not the right way to decide if we should hire somebody. Just looking at last month expenses adding in the new expense and then looking at the bank balance. Right. It needs to be more sophisticated than that. But I think we'll get there. So this shows like some promise, right? And so he goes on, he asks a few more questions and we kind of get the same limitations. Right? It's not actually doing the financial modeling yet, but I think it'll get there. So that's it. That's those are the demos that I pulled out for this week. If we missed anything that you all think that we should be looking at like some cool AI integrations in accounting products, products for accountants. If you're listening, let us know. Email us The Cloud Accounting Podcast at earmark cpe.com. That's cloudaccountingpodcast.com at earmark.com. Or better yet, tag us somewhere on social media because we see all that and send us the link to the video. Anything else this week? David We used up all the time on I and I think it was worth it.
David Leary: [00:52:27] I think just like with every time there's technology or anything to do with tech, you know, our our Congress people. Show that maybe they're not getting it. I don't.
Blake Oliver: [00:52:37] Know. You got to bring us down with politics or at the end.
David Leary: [00:52:40] David So so there was a letter written, you know, it was Senator Chuck Grassley and Senator Maggie Hassan. So Chuck Grassley, Iowa Republican from Iowa, and Maggie Hassan, Democrat, New Hampshire. Ron Wyden, Democrat Oregon, and James Lankford, Republican or sorry, Republican in Oklahoma, they sent a letter to the IRS new chief. Werfel Right. Because they're alarmed that people are going to use AI tools to create tax scams, and they want the IRS to educate taxpayers to how to avoid these.
Blake Oliver: [00:53:17] They don't they don't need AI to create tax scams. They've got plenty of accountants who help them do that.
David Leary: [00:53:24] Yeah, like, yeah, just all those.
Blake Oliver: [00:53:26] Conservation easement CPAs, you know, given the profession a bad name.
David Leary: [00:53:31] It maybe this is just a way for them to like, I don't know it's a way distracts the IRS from its real problems. I don't know.
Blake Oliver: [00:53:38] It's a way to write letters and be, I guess, right.
David Leary: [00:53:41] Referrals should use Chatgpt to write a response letter to them. Yeah. Maybe the best thing ever. That'd be great.
Blake Oliver: [00:53:49] Actually. So I was I was talking with Chris Vanover of Audit Club and he told me that they were using Chatgpt to write comment letters on proposed accounting standards changes. A great example of using it. Write, write, write me that formal letter that I need to send to FASB to get it into the record. All right. So, Giles, it was great. Great chatting with you this week. If folks want to learn more about accounts, follow you online, where should they go?
Giles Pearson: [00:54:25] You can go to WW dot account test.com and to follow me LinkedIn is usually the best.
Blake Oliver: [00:54:35] Charles Pearson on LinkedIn. Awesome. Thanks for joining us Charles. So much fun having you on the show and hope to see you around soon. We're going to go for a hike here in Arizona, right?
Giles Pearson: [00:54:44] We are. It's going to be hot, I think. Not like not like last time, which was which was a bit cool. But I can't.
Blake Oliver: [00:54:50] Believe you're coming to Phenix in June. Like what? Mad Man comes to Phenix in June. That's how I knew that we would get along. Right?
David Leary: [00:54:57] So it's worth budget. Budget conscious.
Blake Oliver: [00:55:01] It's true. You can stay at all the best resorts for nothing. For nothing. My favorite thing is being a local and there's no tourists around. So you go to you can go to the five star resort and pretend you live there, you know?
David Leary: [00:55:14] Oh, all right. Looking forward to that.
Blake Oliver: [00:55:15] Talk to you soon. Bye, David. Bye. Hi, everybody. Hi.