Technologists Randy Johnston and Brian Tankersley, CPA, discuss the latest trends in generative AI, and their impact on accounting firms.
Use the podcast player below to listen.
Transcript (Note: There may be typos due to automated transcription errors. Also, due to the intro to the podcast, add apx. 10 seconds to the time stamps below.)
Randy Johnston, Brian F. Tankersley, CPA.CITP, CGMA.
Randy Johnston 00:05
Good day. I’m Randy Johnston with Brian Tankersley, co host for the technology lab, we are pleased to have you back with us again. And we want to update you on developments in generative AI. Of course, artificial intelligence has been super hot this year. Lots of people are talking about it, lots of webinars and so forth. But we’re trying to just summarize some of the key facts that have happened over the last recent past. And much of this is coming from public information in the quarterly filings by other big players. So Brian, I know you read those things with a fine tooth comb, and you get insights because I, I don’t know, I figure you sit around at night, instead of, you know, watching TV or playing with the dogs. I think you’re reading these reports and ignoring your wife. Oh, you’re not doing that?
Brian F. Tankersley, CPA.CITP, CGMA 00:58
No, no, no, no, no, no, she’s happy girl. But no, we, I’m hanging out with the dogs and the bees a lot these days. And it’s actually pretty good. But you know, when the when the public filings come through, this is the stuff that they’re selling the Wall Street now the and so you the I think it’s an important thing to read when it comes out, you know, I try to read them so you don’t have to, because they talk about how much they’re going to raise prices. They talk about, you know, buy and you get that by how much they expect earnings increase. They talked about how much they’re going to make investments in the new products and lead more or less than in previous years. And they talk about generative AI and the new stuff. And generative AI is an awfully sexy thing to a lot of investors these days. So, and again, these are big companies and so many of them have centralized their generative AI. So while so, you know the first one we want to talk about here is going to be cch Wolters Kluwer. And so they say that they’ve had over 10 years of experience leveraging AI to the benefit of customers, and they’re experimenting with generative large language models. Now, remember Wk has has a lot of stuff in legal DVK. Wolters Kluwer, which is cch Of course, they have what they have. They have legal they have a lot of medical things they have. They have, you know, just a bit as well as tax and accounting that they work with. And so they’ve identified a list here of potential use cases, including conversational search, questioning, answering, writing aid and drafting assistance, translation, document summarization, natural language to code and then personal assistants and chatbots. And Randy, you and I saw a demonstration on a tool that some of your some of your larger firm clients are actually looking at with with Wolters Kluwer, why don’t you talk to us about that a little bit?
Randy Johnston 02:59
Yeah, Brian, I appreciate that. Because, you know, these industries you were just mentioning for Wk, you know, because those of us in accounting tend to think tax and accounting. And we’re used to dealing with Wk, CCA Wolters Kluwer and Thomson Reuters or into it as an example, we only think about the tax and county elements, but health care, finance compliance, legal ESG are all target markets for Wolters Kluwer. And in August, they announced a product called teammate document linker for external audit. Now, this product seems very interesting and it became even more interesting when data Snipper began raising prices. And we have a separate Technology Lab on data Snipper. A very fine extraction tool. But the best I can discern at this point friends, the teammate document linker tool may be as sophisticated for extracting detailed information and the way it maps into audit comp and review tools is very fascinating as well. Now, it turns out this document linker product as some key things that they think are a big deal, they believe that will help Junior auditors save time, they think it will help with the workflow by matching text from scanned images. They think it will, you know, help with higher audit quality and according to the audit quality standards kicking in very very soon. So, you know, this match of the evidence is actually quite important. But then there’s also integration efficiencies that come from this. So they can do a dynamic link between a specific Excel a specific cell and an Excel work paper and the exact location in the evidence document with this product. And it overcomes a lot of common challenges as we see with the product. So I’m quite excited about this, because you and I have talked many times, Brian about these automation tools that, you know, get data out of documents. And we’ve done that in other tech labs, things like the DEXT tool, or things like hub doc, or stage auto entry, these extraction tools are all very interesting. And I think you’ve followed those for a long time. So I might let you give additional color comment on them at this point?
Brian F. Tankersley, CPA.CITP, CGMA 05:40
Well, you know, for a long time, people have been talking about this, this actually kind of this this tool here is I think it’s one of the better examples out there right now, probably the best one I’ve seen thus far. The thing about it is this kind of reminds me you may remember, you and I are of a certain age. So we remember a royal, the Royal Bank of Scotland used to have advertisements that were on televisions and let’s talk make it happen. And I think this is a good example of that. And I think you will hear more about these tools as time goes on, including I fully expect something to be announced at the Thomson Reuters conference, because I think AI is one of those things that everybody has to have one of this year. So you’ll see, I think you’ll hear some of that when in synergy later this fall. But I think this generative AI tool is a winner. And I think it’s a tool that small and mid sized firms ignore to their peril, because of the huge productivity increase that it represents, especially with with junior staff work. And I, you know, I think honestly, the dream of being able to, to automatically match purchase orders and receiving reports and, and invoices together with the documentation and have have that turned into something that’s linked back and forth is a is a pretty exciting thing. You know, I can think about these tic legends, I used to have that were two pieces of 14 column wide that I used to have to do on on a 133 governmental compliance audits. And, you know, I’m, I’m quite excited for the junior staff people that they maybe can have a slightly lower amount of misery as a result. So
Randy Johnston 07:30
yeah, and you know, when I think about this with you, Brian, you and I have the good fortune of working with these development teams, as they, you know, evaluate what they’re going to do with AI tools and where they’re going to appear. And I’m thinking about our Debbie Kay context in Switzerland, as well as our Thomson Reuters contacts in Switzerland that have been behind a lot of this, and how these tools have evolved over time, just like one that’s done some of this extraction and matching work, the validate and validate tools, which we’ve covered in other tech labs as well. But this whole product announcement for cch where they’re talking about their machine learning and other things, a whole lot of the platforms are going to get this including, you know, the CCH IQ product and the one some provisional, you should definitely think about these products. But even cch access tax is going to get some a bot automation. So robotic process automation, machine learning artificial intelligence. I mean, there’s a lot of stuff going on.
Brian F. Tankersley, CPA.CITP, CGMA 08:40
And honestly, I mean, if you think about cch is acquisition of sure prep, it seems like Dave, at one point, it told me that he’d either apply to received a patent on some data extraction from PDF. And and it seems to me that, that I think that IP is probably directly I would expect that to be directly involved in what they’re doing here.
Randy Johnston 09:05
Yeah, and by the way, there’s Brian just said cch, but I know he meant Thomson Reuters. So there’s your prep acquisition.
Brian F. Tankersley, CPA.CITP, CGMA 09:14
You’re correct. You’re correct. I’m I did miss up Thompson and cch there. You’re right. But that I think it’s also indicative of, of the fact that there is an AI race between the competitors. And with that, let’s switch over to Thomson Reuters for a minute, because they have a slightly different strategy that now unlike Wolters Kluwer, Thomson Reuters hasn’t announced a whole lot past they’re sure prep. They’re sure prep acquisition earlier in this year, and I apologize for getting that wrong earlier. But they are building in AI. Now. Thompson has changed their their organization structure in the last few years. And so now now instead of having dedicated development resources For tax and accounting, they’ve combined all their development into a single silo, which means that the legal people are going to get more of the Thompson AR AI earlier than the accounting folks are is what it looks like to me based on based on their disclosures. But they say that they’re going to put, they’re going to build Junior VI and to Westlaw in practical law and something to checkpoint, as well as another tool called high Q that they use. They also have a partnership, they say with Microsoft, 365, copilot the own you know, that’s, that’s, of course, the general AI product from Microsoft, that’s going to be $30, a user that will be on top of your office 365, Microsoft 365 enterprise plan, so it started in the fall. But what they have some ventures in here, but notice that, again, they’re buying AI based solutions to get I think some of these are Aqua hires, and you know, to get some of the talent and to get some of the things in here, but in their lap.
Randy Johnston 11:08
And to your point there, Brian, the case text acquisition, which occurred on June 27, for $650 million was completed on August 17. Now, I’ve been very fascinating on these acquisitions, how quick you’re going from announcement to close. Because, you know, traditionally that had been some months of due diligence. That was only six weeks to complete case text. And it was clear when they bought that because you and I talked about it in June, that it was an AI intellectual property acquisition that they were going to spread over, you know, the organization, but it was also being driven on the legal side. So when you look at the purchase of case text, and sure prepped and CO counsel and you know the evolution of document intelligence from Thomson Reuters. There’s some pretty big things going on. And of course, you and I are aware of some other document management developments that involve AI as well, that are still under nondisclosure. So I’m just looking at this saying Wowsers, plus the Microsoft co pilot 365 announcement or Microsoft announcement saying, We will indemnify intellectual property claims, if you’re using co pilot 365 That gives a little bit of a safety to groups that are using that. So but Thomson Reuters has a lot of other stuff going on too, don’t they?
Brian F. Tankersley, CPA.CITP, CGMA 12:35
Oh, they do. They do. You know, and I I agree with you on the co pilot on the Microsoft identification is a pretty big deal. And you know, I think I think they’re one of the only organizations that can write that check with their mouth, except for maybe Apple just because their market cap is so high. You know, when you have a trillion dollars of market cap, you can you can write a lot of checks with your mouth, but you better be ready to catch them. But looking at Thomson Reuters here, you can see that you can see that some of their newer things that they’ve got are the dynamics, practical law, dynamic search, thought trace, which is an ai, ai driven contract analysis tool, Westlaw precision, and of course, they they put the lead, they have the most recent one in their timeline is sure prep. But like you mentioned that that acquisition, that you that you mentioned in the previous one case, case text. Yeah, was was a pretty big deal. From that perspective, and I think they’re, I think, I think they’re, I expect them to have more tax and accounting announcements, but because they are so heavy on the publisher publishing side, I wouldn’t be surprised at all if the early announcements aren’t related to checkpoint and checkpoint related tools. Because they seem to have more integration on that platform than they do on some of the others with the legal side.
Randy Johnston 14:02
Yeah, and let’s face it, I can’t even remember how many years ago you and I first saw the AI intelligence being put in checkpoint now as demonstrated to us. But for our listeners who can’t really see what’s going on, Thomson have published a timeline of all of their AI development since 1991. And I have also got the Microsoft developments over a common timeframe as well. So we would expect cch and Wk Wolters Kluwer to produce a similar timeline to show their prowess in machine learning and AI over a long period of time, like you said, bro, and it’s part of the story for Wall Street, selling their expertise. And I would expect a similar time line from Intuit to be published at some point in time. So are there other topics on Thomson Reuters We should talk about on AI because I know we need to speak a little bit about into it too.
Brian F. Tankersley, CPA.CITP, CGMA 15:00
No, I don’t think so. But what I would say is that I think that I don’t think we’ve got any slouches in AI and any across any of the big publishers. So I think you underestimate the AI impact in the next few years at your peril. I think it’s I think it’s pretty, pretty interesting times. Now into it. Actually,
Randy Johnston 15:23
before you go on, before you go on to into a Brian, sorry to interrupt you on that. But if you did not see the announcements during the past week or so, it was announced that most of the AI people in Google, most of the AI people at open AI, have changed jobs. And you know, there’s a lot of job hopping that’s going on in this this talents being raided would be my suspicion of that. But I was surprised that over half of the people behind the Google AI strategies had been hired by other companies, according to press announcements this week. So I just want you to understand that these big companies have some great AI talent. But there are just a relatively few superstars that can make a huge difference here. So punt into,
Brian F. Tankersley, CPA.CITP, CGMA 16:15
yeah, and they’re there, they can get whatever they want right now in the market, because Wall Street is just way too excited about this. But into it, you may remember last year forced a lot of folks off of their payroll products. And they changed up the payroll products. So they actually took the old pay cycle product on QuickBooks online payroll, and they consolidate that into a new Intuit online payroll platform. And part of the reason they did is because Intuit, you’ll recall, actually has MailChimp and, and QuickBooks and TurboTax, and QuickBooks payroll, and you know, just a number of different businesses, that they that are credit, karma is another one in there. And so Intuit strategy seems to be a little bit different. And they seem to be using it more to the end of incorporating Consumer Finance and Business Finance, and related models into it. So it’s, I feel like when I looked into it, they say they’ve got information on 10 million small businesses and customers through QuickBooks and MailChimp, and 93 million consumers through TurboTax and Credit Karma. And it seems like they’ve almost decided that they want to be a credit bureau, instead of instead of jumping out and, and doing and necessarily solving the problems directly. So they’re talking more to Wall Street, about that capability, as opposed to, as opposed to talking about how we’re solving problems for practitioners. Okay.
Randy Johnston 17:53
So Brian, I think it was pre pandemic that you called the shot around the focus of Intuit on all of this Credit Karma, you know, finance piece, and less on QuickBooks, and Lacerte and pro series, but I will give into it a little credit here, they have certainly been pulling along their tax and advisory pieces. But you know, these most recent developments, kind of resolve a lot of this strategy in this consumer centric approach. And, you know, through the years into it a swung towards consumers and away from professionals. And then back to the CPAs. I can’t even count how many times I’ve watched that cycle. But right now, with I think that they are favoring this consumer side, and not as much the professionals.
Brian F. Tankersley, CPA.CITP, CGMA 18:51
Randy Johnston 22:48
Yeah, so Brian, just to give our listeners some timing on this, because again, we don’t know exactly when you listen to these things. But the Jin O ‘s announcements were made on June 6 of 2023. And the Intuit assist announcements were made on September 6, of 2023. And Intuit supported this with a document that they called the Intuit data and AI fact sheet. And in that they said Our strategy is to be an AI driven expert platform. And they went on to say their foundational was the data that their customers have entrusted with them, which includes 60 petabytes of data 500,000 Customer financial attributes per small business, 60,000 text, financial attributes per customer, and 20 billion transactions imported from financial institutions annually. And then they went on to say that they’ve done 810 million AI driven customer interactions per year, we have 65 billion machine learning predictions per day, with 25 million conversations processed with natural language processing per year, with 2 million models running in production per day, with approximately 900 AI machine learning and data science, US Patent assets. That’s pretty interesting. And you know, they said that they declared our strategy to be an AI driven platform five years ago, which kind of aligns with this whole consumer facing thing, but they weren’t so clear in their messaging. I thought that was a pretty interesting fact sheet.
Brian F. Tankersley, CPA.CITP, CGMA 24:34
Randy Johnston 25:28
Well, Brian, I know we’ve got a little longer on the tech lab, and then we commonly would, but we thought it was important that you understood the public announcements from your major suppliers of cch Wolters Kluwer Thomson Reuters and into it. So much more to say, but we thought we’d give you a kind of an update, which, you know, summarizes so much of what’s happened in this AI frenzy around us. So, Brian, any parting thoughts?
Brian F. Tankersley, CPA.CITP, CGMA 25:59
No, I think it’s an interesting time. And I think you I think, I think you’re going to see that I think that AI is going to make changes to the profession and how you work and how I work sooner rather than later. So I think this is going to move a lot faster than the, say, the dosta Windows transition that many of you made, or the windows to cloud transitions, some of you may, I think this is going to move a lot faster. And I think it’s going to require you to be a little more agile, as you respond to these, these changes. Because, you know, let’s face it, the talent market isn’t going to get any easier anytime soon.
Randy Johnston 26:36
Yeah, and you know, my one reminder, I’ve been watching a lot of, you know, AI expert wannabes in the market, showing you how to do things which are very fascinating, but ignoring the privacy and hallucination issues of these models. And to my knowledge at the point of time of recording this tech lab for you. No vendor has solved that those particular problems. They have got private databases. So the privacy is a little less than a shoe, but still knits and a shoe. So anyway, was such a pleasure to have you with us today. Hopefully this has given you some things to think about and things to innovate with, and things to be cautious with and frankly, ways to serve your clients better. We’ll see it in another technology lab. Good day.
— == END ==–