Balancing the Books with AI: Innovations in Accounting Practices

John Littler: Well, welcome and good afternoon everyone. Excited for you all to join us today. Before we take a seat, I did want to take a moment and introduce Sid. Sid is the CEO and co-founder of Docket, an AI powered accounting automation software platform for business owners, accountants, and CFOs. Sid's extensive background includes being the designer and architect at Silicon Valley companies including VMware, Oracle and Whirlpool. We're excited to share the time with you this afternoon. Sid, have a seat.

Sidharth Saxena: Thank you. Nice to meet you.

John: All right, Sid. So, we've had a chance to meet, catch up. AI topics, throughout the event here. Most interesting thing when we thought about this, I've got a team and we're out there talking with accountants and small businesses, and all we keep hearing is "AI. AI. AI. AI." So, what was the impact on ERPs, on all of this cloud evolution and haven't they come a long way since the time of Lotus?

Sidharth: Absolutely. You know, interesting that you mentioned the word Lotus. I could go all the way back to VisiCalc, which was the first viral app in 1979 when it got launched. VisiCalc literally pulled the PC sales single handedly. That was how personal computer revolution really started. But coming to the topic of ERPs, these products have to be very cautious in how they evolve, because guts of a financial corporations or guts of an enterprise runs on their software. So, they are slow movers, slow adopters. Even cloud as a technology, they were laggards in adopting AI.

You know, they are now scratching the surface of it, but reality is, an ERP system owns the guts of some corporation’s financial data, which means they have an iron grip of their customer. These customers are not running away to the new shiny tool anytime tomorrow, which means the scope of innovation is really slow. Innovation in accounting software actually has happened in a piecemeal way, because general ledger never innovated. So you got a cloud company doing bill pay. You got a cloud company doing expense reports. You have a cloud company building e-commerce later integrations. So there was piecemeal scope of innovation and now slowly ERP companies are catching up to AI technology.

John: And I know you talked a little bit about when Excel was there. There was that one step before after Lotus, was Excel. And what type of impacted Excel and how is AI now kind of overtaking?

Sidharth: Absolutely. That's again, going down the memory lane. So after VisiCalc, Lotus, when it came out, one of the first innovation that Lotus did was actually bring in automation. It was the first spreadsheet program that brought in macros. For the first time, at the press of a few clicks on the keyboard, you can automate a sequence. Imagine, you know, accountants actually powering automation in personal computing. But, Excel was actually the first program that was designed for graphical user interfaces from the get go.

Innovation in user experience and design actually outsold the product very, very rapidly. There was a time when VisiCalc was hoping a million dollars in revenue, and they earned $59 million in that year. But when Excel came out in 1985, all of that changed the game. So now, accountants actually had a tool that was not only powerful, running on extremely cheap personal computers, but also was user friendly.

So, whenever computing evolves, you will see this common thread that first comes the infrastructure. In that era, it was a personal computer. Then comes the first generation of automation software. And then, the user experience evolves it to a level where it becomes easy for people to absorb that technology. We are seeing a similar approach with first cloud becoming mainstream. Late ‘90s, early 2000s, the infrastructure is now there to absorb tremendous amount of data. Then the technological evolution came in artificial intelligence—Saas, cloud software. Large number of people are now used to using cloud software on their desktops and mobile devices. But finally, we're going to see another layer of innovation from a user experience perspective. It becomes easy to absorb all of this innovation in AI.

John: So that brings us to today, right? So why is real-time accounting so important today?

Sidharth: That's another very good question. You know, nobody ever said, "Give me my books a month later." Everybody wants their data now and by now, literally now, and if you think about it, most business software actually are real-time. In real time, you can see the ad performance of your marketing campaigns. In real time, you can go to your Salesforce and look up your sales funnel. Why are books still getting delivered late?

When you think about it, you know, the process of book closing, because it's such a sensitive process and because of this piecemeal innovation, data is spread far and wide. You have some data in accounts payable. You have some data in, you know, travel and expense ledger. You have some data in point-of-sale system, payroll systems. Somebody has to assimilate all of that carefully and make sure that proper processes are followed to actually close out the books. In that process, what happens is human gatekeeping. Slowness of these multiple technological systems really delay the process of closing the book. This is where, you know the way fragmentation in this industry has happened has really slowed down things.

You can either get accuracy or cost or speed. You can pick two of these three. If you want all three, you need a humongous amount of resources. Accountants want accuracy, so that happened. Businesses want low cost, so that happened. And speed became the roadkill. Now, because of evolution of AI and powerful software speed actually can also be achieved. So we are at that cusp of innovation where real-time accounting actually is going to become possible.

John: Interesting. So based on everything you shared, is this agnostic to industry? Are there specific industries where it's really taking off? If you can give us kind of a guide there, where are you seeing the AI being impacted in the county and what industries?

Sidharth: Absolutely. So accounting and finance basically touches every industry. You know, every business, the root of their success lies in how nicely and better they manage their financials. If you are a smaller business, until now, you did not have access to good technology. A big corporation can have hundreds of accountants, can invest millions of dollars in an ERP system, and they can, by brute force, get stuff done. But when cloud evolution happened, medium-sized businesses now could do the same thing which large enterprises were able to do. Because of AI technology, now accountants will have tools which can scale them. And they can now service lots of smaller businesses, which previously could not afford accountants or their skill set.

One of the common reasons SMBs fail is because either they run out of money or they're not able to manage their business well. Expertise of accountants can actually really help save a lot of these businesses. So, industries like, for example, hospitality, industries like e-commerce and retail, where, you know, managing your inventory properly, you have real-time visibility into that data. Companies which are fluid workforce, hotels, if they have real time visibility into their payroll. Companies which are, you know, project centric, if they're consultants, is the workforce being utilized in a proper way? If you have real-time visibility into your KPIs, you can actually run your business more efficiently.

John: Hmm. So, let me ask you, we have the AI, ChatGPT. Do you see ChatGPT closing books. Is that realistic?

Sidharth: If you can avoid the hallucination problem, you probably can close books. But ChatGPT is a very specific kind of AI algorithm. It's a large language model, which in ChatGPT's case, they're a paying company, open AI has travels and absorbed internet's vast troves of information. But it is really trying to please you as the person who has asked the question. It is trying to predict, what does this person want me to do? And in that process, it will hallucinate, it will make up things. You don't want your balance sheet to be made up in that manner.

So one of the evolutions that's gonna happen, and already happening in the industry, and some of that is something that my company also focuses on. There are gonna be these two flavors of AI technology. One is precision AI and second is generative AI. Precision AI is your traditional AI and machine learning algorithms that are running on workflows—your expense workflows, your revenue reconciliation workflows, your transaction coding categorization workflows, your payroll accounting workflows. Precision AI can extract information, process it, and with 99.9% accuracy, tackle over 80% of items and this is what we are seeing as a pattern.

There's an 80/20 thing going on, where precision AI can really remove 80% of hard work. But what about that other 20%? That's the last mile, meaning once all the data has been absorbed, once all of the transactions have been coming in and some of them are remaining, then the humans need to come in. Which is where an accountant, you can expect doing human Q&A to their clients, asking, "Hey, what is this check for? Who got paid?" But, the moment humans come in, technology slows down. That's where generative AI can do this last mile, 20% work for humans, where things can be flagged up.

And imagine in a mobile app for a business customer. Now, in an interesting way, some of these things show up and they can quickly just, through voice, respond to questions. So essentially, when you combine precision AI, with a large language model like generative AI, you actually can achieve real-time accounting now. So, to extend your answer on, "will ChatGPT be closing out balance sheet?" Maybe not ChatGPT itself, but a product that has both of these kinds of technologies combined together in an innovative way could actually go line item by line item in the balance sheet and actually close them.

John: Oh, interesting. So with that, where do you see the accounting world headed? I mean, one of the things we talked about, one of the big blogs that is getting a lot of hits are accountants are nervous that jobs might be eliminated or whether the case might be, but it could actually enhance their roles.

Sidharth: Absolutely. I was saying, you know, if your time gets freed up, you know, you become a happier person. You don't feel like, "Hey, somehow my time needs to be consumed by stuff." Accountants come in this profession, not to do spreadsheets. If you are in this profession to, you know, manage spreadsheets, please raise your hand and I'll stand corrected. But most people come in this profession to actually help businesses succeed. Businesses and small businesses are the foundation of America. They are the foundation of capitalism and a successful business actually, you know, can contribute a lot towards small business and in their local neighborhoods.

An accountant who can scale themselves, currently, what we see is a bookkeeper can assist maybe eight to 10 businesses through their effort, but with all this technology, they can now provide their assistance to 30 to 40 businesses. Three times to four times their business, which is tremendous for them and their neighborhood and their customers. So number one, we will unleash a lot of support to small businesses. And by the way, good accountants, they are not making them anymore. You know, really good accountants are retiring and kids are not joining this profession in the volumes that the business formation is happening. So, there's a big shortage of accountants happening. What's needed is some technology to take the grunt work out, so the advisory and assistance, higher value services can be offered to more businesses. And that's what this AI technology is going to unleash. But I want to extend your question on real-time accounting over here as well. That how does small business really, you know, get help?

Let's say you have a point-of-sale system as your revenue platform. Today, a typical small business accounts for their point-of-sale revenue at the end of the month, which means for almost, you know, 15, 20, 30 days, they don't have liquidity visibility into their books. But, imagine if accounting of their revenue was happening daily. Point-of-sale systems generate these cumbersome reports, which accountants use to input data. But, accounting resources are limited, so it's not happening as frequently as it should be. Again, automated systems, AI systems, they can absorb this data in real time and make your income statement live, so you can see your net income at any given point of time. So all of this is going to make businesses actually more successful and a lot of unnecessary headache and grunt work is going to go away from accountants.

John: So there, to that question, will there be no accountants in the future or…

Sidharth: I do not see any scenario where accountants are not needed. The most important area where accountants, you know. Once you do things like, closing books, you understand the growth drivers of all the businesses you support. You really become an expert on what is making your customer successful or where are your customers failing. Those are tremendously powerful experiences that allows accountants to help small businesses. So businesses will always need these advisors. Even myself running a company, a tech company in Silicon Valley, I surround myself with advisors of various kinds and somebody who can look at my balance sheet and in one glance they can say, "Hey, you know, you got to show up more capital. You got to do X, Y, Z," you know, that kind of insight while I'm trying to really build a business.

I may not be looking at some of these important characteristics. You will always need those knowledgeable people. So, it's wrong to say that accountants will go out of job. I think of accountants really as business advisors. Every business will need advisors. And if you want to extend that, you know, a medium-sized business, if you go above that SMB level, one of the characteristics of medium-sized businesses, they have departments, but is the marketing department head, does she have access to real-time financials for her department? Does the sales department have? Does that person have real-time access to all the commissions, all the information that, you know, are there within budget? Even that information is not available to successful medium-sized businesses. Real-time accounting will unleash that and the internal finance teams, which are responsible finance managers, which are responsible to really assist and help every department in an enterprise succeed, now they're going to be able to unleash their superpower and go help keep marketing in budget, sales in budget.

And if the business may not meet the net income or even gross operating income milestones for the month, maybe today's 18th and you see milestones not being met, maybe you can run a marketing campaign. Maybe you can incentivize your sales team in real time so that for the remainder of the 10 days, you can actually meet your milestones. This is the power that, you know, real-time technology is gonna bring in accounting and finance. There's an additional layer of financial, you know, capabilities that are gonna be unleashed, which we can probably talk more.

John: So, when I think about it, over the last 12 to 18 months, we're seeing the client accounting services or client advisory services growing, AI growing in accounting, where is that aligned now? I mean, it's really positioning accountants to work with their clients a lot differently.

Sidharth: Absolutely. And, this is where accountants will have to really decide on what excites them more. You absolutely have the domains of auditing, you have domains of tax filing, you have domains of client advisory. When you think about bookkeeping and book closing, those are areas which will immediately benefit from AI technology. If you are in the business of closing books and running a bookkeeping company, then you have to make a decision. Do you want to scale your customer base? Or do you want to go upstream in the kind of services you want to provide? You don't want to be in the business where you have been shuffling spreadsheets and trying to close books for tax time and then your job is done.

Actually, the accounting job starts after the books are closed. But somehow majority of professionals in this area, you know, because it's such a monthly repetitive process, it takes so much time. One month's books are closed, then the immediately next day, the book closing work for the new month starts. So there's no breathing space. This technology is going to give people a breathing space. They can use it to expand the business or go upstream. So advisory services will be a great area for people to experiment, and again, generative AI can help them. You know, generative AI technology can understand data from a closed book and provide advice on that, "Hey, you have a lot of inventory that you should optimize a little bit." There can be standardized reporting that accountants can now start sharing with their customers. So in advisory services, there will be a lot of innovation coming in.

John: Okay. So speak to the pivoting peaks, because I think that is really, really important. You mentioned, and he gave me an example earlier about hospitality, where they can actually look at, you know, the data that they're seeing and make much better business decisions. Can you elaborate a little bit more on that?

Sidharth: Absolutely. Imagine you are a Courtyard Marriott business owner. You are successful. You will not stop at that one location. You're a franchise business owner, now you know the game. You will buy your fifth Marriott, the tenth Marriott. You want to have an empire. When you are operating at that scale, even cost of simple things, like get supplies, linen, now you have to be unit economics driven. But if your books are getting closed 15 days, 30 days after the month end, you're already behind on unit economics. Your teams are running a little bit blind. Your general manager of one location does not have real-time report based on which they can change things. If it's summer, let's say June, and electricity bill is gonna be a 1,000 bucks, and if you know that by end of July, you can't raise rates in June month once June has ended.

So, all of these kinds of things where you have unit economics for your utilities, your electricity rate per available room, your linen cost per occupied room, these kind of ratios, unit economics. Imagine running live on a dashboard, chart of account by chart of account, category by category, multi-location consolidations. Consolidation is a big problem for businesses, which have either multiple locations, subsidiaries. Imagine consolidation happening in real time. So now, not just for one location, multiple geographies, multiple entities, you can have unit economics power dashboards. And small businesses which adopt these technologies are gonna be the winners, folks who don't will stay in a smaller business.

John: Absolutely, incredible. So, I know we have just a few minutes left. So, I know there might be some questions that come from the audience. If you do have a question, please raise your hand. We have mics. We'll make sure that you can be heard. Question?

Audience member #1: Can you talk about your technology and what you've built?

Sidharth: Absolutely. The question is "Can I talk about my technology and what we have built?" So my company is called Docket and we build accounting automation technology. Our purpose and mission is to make accounting real time and the way we do that is exactly what I described a few minutes back. We have AI models that do precision, you know, workflows. We have workflows to do revenue accounting bill pay, expense reporting, and then we have workflows that actually close the books, do the reconciliation, and finally, we are applying generative AI in that last mile human to human collaboration area so that books can be delivered in real time. So Docket has precision AI and generative AI combined together.

John: Amazing platform. Amazing platform. What else? I thought I saw…

Audience member #2: Since you said this, advisors are business advisors, right? Wouldn't an AI do a better job at that than a human could do? An AI can analyze way more data faster and more efficiently than a human could ever do. Just a fact.

Sidharth: Very good question. And again, I'll repeat the question. So he said "Accountants are gonna be business advisors. Wouldn't AI be a better business advisor?" That's actually a very good question. And the reason is absolutely, AI will know, let's say, about the entire industry vertical. Let's say hospitality as a vertical. AI knows what are the key metrics, what are the key ratios, occupancy rates of hotel, revenue per available room kept, KPI metric. It will have generic knowledge, but when it comes to knowing John, and I know John doesn't like to work on Fridays, you know, I can tweak some of the information that John needs, which works for John. No AI technology, or at least not that I know of, can understand how humans are emotionally connected to their business. Or maybe what could be going on in the business owner's life. Maybe they're expecting a kid. Maybe, you know, they are thinking of expanding the business. If you're thinking of expanding the business, maybe you need to do better liquidity management.

Human intent, until technology can read human intent, only humans like to only talk to other humans. I could be playing golf with John and say, "Hey, I want to buy these five locations, you know? What should I do?" I'm not going to ChatGPT and type to buy five locations, what should I do? So I think that's where the difference comes. And that's why that human connection is something that accountants, they're already doing that, but now they will have a greater opportunity to connect with their customers on that human level.

John: Is this along the same lines or is it somewhat different?

Sidharth: I think it's on the similar lines of the question on the screen. If I read, "What could a human provide in terms of analysis accounting that an advanced AI specifically made for the task wouldn't." This is good, you know, the more specific you go. Was this your question? Fantastic. And again, the more specific you go and the better intent you can provide to technology, the better the results is going to be.

John: Got another one popping up. You mentioned 80% of accounting tasks will be accomplished with AI. More specifically, what are the initial accounting tasks that AI will assume?

Sidharth: So, data absorption and triage of the data and matching of data. Computers are way faster at that. You can scan receipts all day. But computers will extract data out of all those receipts in a few seconds. Bank transactions are flowing in. Humans will take 20, 30 seconds to match a receipt with a transaction; computers are going to do that extremely rapidly. So wherever data is involved, and again, this is, I'm talking to context of book closing process, computers can close books much faster if data is available to them. So, these are the 80, this is unfortunate reality. 80% of a bookkeeper's job is stuff where machines are actually really fast and good, but we haven't given them the tools, yet. Now, for the first time, because of all this technological evolution, there are tools coming their way, where point-of-sale revenue can be recognized daily, where all of the expense documents can be matched with transactions, instantaneously. And every ledger of the balance sheet is being looked at continuously.

So reconciliation is happening continuously, not 30 days after the month has ended. So these are the tasks, which are precision AI, where these micro machine learning models are working on financial workflows, can do with great accuracy. This is almost the 80% work. And then the remainder 20%, the last mile work is where humans need information from other humans who got paid.

What is this check number 1-2-3-4 or some customer have to respond to 15 transactions, which have been flagged by the accountant. Today, most accountants send a spreadsheet that, "Hey, John, these 20 transactions, can you help me understand?" Now, John will have to open up the laptop, open up the spreadsheet program, start typing in. A busy business owner, which have 10, 15 locations to manage, they don't have that kind of time. So this is where, you know, some of the technological evolutions will help.

John: And I think just in the last minute we have a chance to talk to accounts, small businesses. So it sounds like it's really going to be helpful when you have September 15th deadlines happening Friday that the client and the account will be all on the same page.

Sidharth: Absolutely, last minute catch up is a tremendous use case. You throw in shoeboxes of receipts at the system. Within few minutes, all this data will be pulled in. You throw in, yeah, two, three years’ worth of data or transactions in the AI system. Within half an hour, it can categorize based on industry vertical knowledge. This is how Docket's system works. So within half a day, most of the grunt work is done. And now the remaining few things are left to tie together and your tax filing deadline will be met.

John: Amazing. Win-win for everybody.

Sidharth: Absolutely.

John: So Sid, we are up on time. I can't thank you enough. It was just great to get to know you, learn exactly how your platform's going to help businesses out there, and working with accounting firms, which we specialize in as well. So, we can't thank you enough for all the time.

Sidharth: I really appreciate TriNet inviting us over here.

John: Yeah. And if you liked today's session, please give us five stars, then we'll go through that. But we all thank you very much for joining us this afternoon. We really appreciate it.

Sidharth: Thank you.