How AI Will Transform Accounting: A $100B Opportunity Explained

accounting is one of the areas where the software that's being used is like Reagan Era software the consumer and business demand for these professionals is increasing your overall workload is remaining the same or increasing but your number of staff that are hired and retained to perform that work is either remaining flat or declining if software could take away a lot of that manual stuff that should ultimately reflect in in cheaper prices in Faster service and generally positive consumer experiences [Music] so when people talk about jobs that are disappearing they frequently talk about things like coal

miners or like you know physical film editors or even like door to-door salesmen but one profession people don't actually talk about that is disappearing as well is accounting like why is that like this is something I know you guys have been looking at like you know why is this becoming like a less common field yeah and actually the stat is I think 75% of CPAs are set to retire in the next decade um um and uh the new rate of CPA graduates is not nowhere near going to match that and I think we'd seen another

stat that uh somewhere around 16% of the number of accountants had dropped between 2022 and 20 uh 2019 to 2022 U which is a pretty big drop for the profession um I i' probably point out two big market trends um one so yes there is this uh decline in the number of people who want to become uh accountants um or or you know generally be in the accounting field which could include bookkeepers and other folks that work in in corporate accounting roles the amount of accounting work and the complexity of tax has only increased over

time so the consumer and business demand for these professionals is increasing so as a result you're end up you you're seeing these firms have a real Jam in terms of being able to meet the demand and even more pressure is getting put on professionals so that's what I kind of highlight as point one point two is even with the um decline in the number of professionals and increase in work it's still a huge huge Market um and so I think the the rough stats there's about about 1.5 million accountants and Auditors and then if you

broaden that out to accounting related positions whether it's in corporate finance departments bookkeeping payroll Etc there's there's over three million um individuals and that's probably well over um hundred billion dollar of of spending in wages so it's huge um and and that's a lot of hours of work that need to get done also one way or another and so I think this is a really interesting Dynamic that creates a really interesting wi now that you know people may not know about because it's not a physical industry um but uh it's definitely something where there's a

lot of repetitive manual work that's going on behind the scenes at the end of the day it's a professional service business and it relies upon client service and so to the extent that your overall workload is remaining the same or increasing but your number of Step staff that are hired and retained to perform that work is either remaining flat or declining there becomes just pressure uh it becomes harder for each individual accountant to provide the same level of service to their customer and it's something that affects everybody like as you guys kind of point out

in your piece right as been Franklin has said everyone's you've got death and you've got taxes so it's it affects every single person everyone's got to pay their taxes yep y you mentioned like Professional Services obviously this is something that you guys look at a lot in terms of industries that are ripe for a disruption so what makes accounting like particularly like a good subject to kind of tackle we generally look at Professional Services verticals that are ripe for AI under the lens of okay where are there a lot of highly paid people who are

doing in many cases repetitive work um that probably could benefit from smarter Automation and I think accounting has that in Spades um I think one of the things that makes accounting a little bit unique is relative to let's say law which is another vertical we spent a lot of time looking at um law is primarily uh a text dominated business so with the current capabilities of llms which are text in text out um they're very capable at handling a lot of the querying and research and inputs and outputs that the legal profession requires accounting on

the other hand as we all know is extremely quantitative in nature um you're dealing with numbers all the time and so what's interesting about it is while it has those characteristics that make it make you want to automate it um it's been a little bit slower going relative to its peer in in legal services so you know in in legal AI there's this company called Harvey that many of the people listening to this will be familiar with um has grown really rapidly uh there isn't really like a breakout Harvey in accounting just yet um and

we think about that and talk about that all the time and we wonder okay is is that a function of the fact that maybe the models aren't as well suited to accounting on day one or is something else at play or are the buying behaviors of these two verticles much different than we would have thought you know when we examine markets and look for investment opportunities we often start with something we call a y now which is you know simply put why is now the right time to invest in this in this category and a

lot of the demographic trends that we've already hit on the the pace of retirement um coupled with the pace of lack of replenishment we can call it you know creates a lot of urgency to fill gaps fill labor gaps these firms from the very top down are feeling tremendous pressure to fill these gaps not just with additional humans but but really with software and so um I think that makes you know relative to other Professional Services verticals it makes this one like a really compelling one to examine closely so when you guys first wrote this

piece about accounting in AI you guys heard back from like a lot of people in the audience people who read it who are interested what sort of comments were people sending you guys like what were people most interested in or most worried about we heard that they they were in resounding agreement I think of a lot of the the themes that we pulled out um they're feeling the pressure internally they want their firms to be more efficient they want their people to be happier um and you know there's this change of moving from uh being

a a doer to a reviewer and we've talked a lot about that in the past um and these firms want you know want their people to be in more of the review seat rather than the due seat um and that also ma matches some of the the labor Trends we we pointed out earlier um and so one of the themes actually one of the firms we talked to it's a top 20 firm they were ready to spend up to $500 million either building or acquiring software um because they think this is so important to to

the direction that the industry is going in so people are interested what are some of the hurdles you know in terms of like implementing a lot of these AI Solutions into Professional Services look I think it's like very high stakes work that that has very limited tolerance for incorrectness so whereas the more creative Fields might benefit from the hallucinations of llms and you can create cool outputs with kind of anything that a model gives you back uh the contents of your tax return and your financial statement audits just have to be 100% accurate so I

think you know while firms are enthusiastic and eager to buy things and be forward leaning they are very conservative about what will get rolled out and how quickly it will get rolled out just to make sure that they're delivering a gold standard that they always have the second that we talk about a lot is kind of an incentive misalignment or potential incentive misalignment or business model hurdle which is uh the accounting industry in general has tended to run on uh the business model of billable hours and so if you're coming in and you're trying to

sell software to someone that uh the end result of which is that they can perform their job in 20% of the time it used to take them if they're being intellectually honest with their customers then that should in theory reduce their billable hours quite a bit we think this is like a short-term hurdle we do think that the way this Market will Trend over time is more towards a fixed Fe engagement and that just the competitive pressures of supply and demand will make it such that if you can deliver the same work product at a

fraction of the cost then you actually should be billing less this has like huge implications about who you're trying to sell the software to we talk a lot when we're dilen in companies about who's your buyer um if you're trying to sell this AI software to like an individual contributor you know they might be a little bit skeptical they might say oh uh I see what you're trying to do here you're trying to make it that I'm Obsolete and uh even though they don't want to do all the manual stuff um as much as they

ever did you know it might they might think twice whereas if you're selling it to their their boss or their boss's boss or their firm leadership or something and they see like the clear Roi opportunity in the business you know uh maybe it's an easier sell so the llms can actually do a lot within the field of accounting but what are they actually really good at as opposed to just something where they're doing like a perfunctory job if you stop to think about what llms are really good at uh data collection and ingestion would be

like the canonical example um and what that means in the context of accounting is you know if if you sit in this client advisory Service practice of one of these firms you're asking your client to give you an export of their general ledger copies of their bank statements receipts PDFs of various contracts they have with various vendors like all sorts of things that exist in different formats and different systems uh and basically translate that into a journal a journal entry into an accounting report um LM are are really good at at making sense of unstructured

data so this is where we've seen them sort of most pervasive in the early Innings of this movement um I would say the same goes for research um llms again excel at text in text out so if you're a practitioner and you want you know an opinion on how you should treat a certain transaction or you want to look something up in generally accepted accounting principles sure you could do that on Google or you could do that through one of the more Legacy software tools but um using LM to do that you can get a

really robust opinion um you can get citations back on you know either like President cases or um you know different ways to think about it so I think like those first two jobs to be done are where we're seeing the majority of the activity we're slowly working our way towards the latter two jobs to be done um so report generation and and filing I think um is a little bit more tax specific um and we're starting to see some Founders kind of coales around rebuilding the tax engine using Ai and then um we're still in

the very early days of of client advisory servic and kind of summarizing all the learnings so accounting software must have existed in some form what has it looked like historically and like you know what is the big gap in terms of like you know what was' been missing and like what we can do now most companies that had success in the accounting ecosystem really sold through accounting firms and what that meant was you basically got these large firms and small firms to be your allies in distributing your product and that would be like the bill.com

of the world would be like Sage intact any sort of general ledger Payment Processing type product um that the accountants could kind of ask their endusers to use to make their lives easier we still think that's a awesome and effective way of Distributing products but we're excited that now is the first time uh that a that a software company can build a very large business just selling two accounting firms as well if you looked at the last generation of um venture-backed attempts at software they actually have to have like 80 accountants on the team sitting

in like you know in a room and turn turning it into the into the code and that was expensive and costly and and it didn't work now we're at a point where llms can do a lot of that work without needing to hire an entire room of accounts um and and Engineers paired with them to to CAU you know to turn the um the accounting code into software it's interesting that there are so many AI developments within accounting because I feel like accounting is such like an old profession it's been around like forever because people

have always needed to keep track of like their money and arguably like you know the the technology that's gone with it is also super old so it's funny to kind of make this jump from like you know paper and Pen to like you know all of a sudden like AI is doing everything tax exactly yeah but it's it is also wild that accounting is one of the areas where the software that's being used is like Reagan Era uh developed software even if you you know if you use one of the personal income tax um calculation

softwares those have been around forever and they've gotten marginally better um and you know I I uh did some contract worker into it 10 years ago and they were still like people buying the CD ROMs um and and using those to file their taxes um so the the software really hasn't gotten better both on the consumer or you know the end user side and for the accounting professional so I think there's a real a a real time for this to to change we as consumers should all be excited about this because uh we already talked

about how this is a profession that runs unbillable hours and so all the rote and manual tasks that were describing whether it's made clear to you or not like you're paying for that in some capacity um and you're paying someone who's has a specialized degree and you know a commenor at hourly rate uh to basically do that type of work they also do a ton of Highly value added of things on top of it um but you know if software could take away a lot of that manual stuff and allow them to get to the

good stuff um that should ultimately reflect in in cheaper prices in Faster service and um generally positive consumer experiences