Robotic Process Automation In Finance

So in the past your accountant will have to search for different keywords, filter and then you can categorize them in another column. Now it's just a matter of pressing three buttons less than a minute. Everything is classified nicely. So with automation you're going to save time, you're going to reduce the human errors and then definitely you can focus on more value added at your work.

Hello and welcome. As we know, AI is changing the world of small business and that goes double in the world of finance. And it's probably happening in a lot more ways than you realize. I had a great call with one of our clients recently, Shannon Schmidt, and she shared how she's using AI in a 20 person accounting practice and how she's taken. It was just one task that she ran through some process automation and she saved over a thousand hours of staff time. So that's basically a part time team members hours for a full year. She was able to save by applying some process automation and it makes you realize how much opportunity is here. And we were having a little bit of a discussion with Shannon and she mentioned one of her secret weapons. Who is Daryl or and he's the founder of skybots, which is a consulting company that specializes in robotic process automation, RPA and also AI as well. Basically what they do is they take a lot of auditing and tax finance secretarial tasks and then let the machines complete it. So I'd love to explore this topic today. So first things first. Just wanted to welcome you to the call.

Daryl, hi. Yeah, thank you for inviting me for the event today.

Oh, perfect. Well, maybe just to start because I know you've got some different areas of specialty. I'd love to explore the area of rpa, the robotic process automation. I know that's where Shannon got started and I thought maybe that might be a great way to start. Just to kind of have you explain a little bit about what that is and how it differs from AI and some of the other ways that you automate process.

So Robotic Process automation in short stands for RPA basically is building a software robot on your computer that can automate your task for you. So it is a low code software that you can give simple instructions and the robot will follow your instructions to the team. So for example, you may tell the robot, click on these three buttons and the robot will click the three buttons for you. If you tell the robot, copy this and paste it there, the robot will do the same thing for you. And the advantage that a robot has is that the robot can repeat the process Many times. And robot can do this much faster than you. So if you have a lot of copying and pasting of data, instead of you doing the manual journal entry, the robot can actually paste, automate this very easily for you.

Yeah, so when we think about, I do a lot of work with process and getting businesses very clear on core tasks, and one of the first things we do is around documentation and we, we capture that. Kind of sounds like you're taking that and programming it then so a machine can complete it.

Yes, correct. So the advantage, rather previously you had to get your IT colleague to help you to write some programming script to automate your work. But with rpa, you can actually do it yourself because how you give the robot's instructions is very simple. Just drag and drop, tell the robot what to copy, what to do, and then the robot will just execute the task for you. So because you are the end user, you will know how to build a user centric solution and then you can actually use RPA to actually build a bot to do it for you.

Yeah, perfect. So it seems like it's very specific use cases where certain things happen. As we're starting to see kind of the developments in the space with AI, which is able to make decisions and do some extra thinking, how are you starting to see that change the world of RPA?

Okay, so when ChatGPT came out, it has actually changed the game of RPA too, because in the past I would do everything in RPA to automate the work. But with ChatGPT, if you know just how to copy and paste the code from ChatGPT, you can actually automate a lot of things. So my class, I teach people how to prompt, for example, how to get a VBA code to perform a certain task in Excel for you. So in simple English, you just need to tell the robot what you want to do in the Excel file. And then I teach them how to copy the code, go into your Excel file, go into VBA and paste the code in. And when you run the robot, the macro will automate the things for you. So the accountant of the future needs to learn how to read code, because once you know how to prompt properly to tell the robot to get the, to tell ChatGPT to give it the correct code. It's just a matter of copying and pasting and run. And then the automation will help you to do the work for you.

Yeah. How do you see where this might happen in the future with, you know, so as AI agents and things like that start to develop and you know, Just being. I mean, we see it with ChatGPT, we can have just natural conversation and get it to do complex things. And they've just launched the latest version of you ChatGPT, which has some reasoning built into it. How do you see that changing? You know, I think it's still really important to understand the core process because then you need to ask the machine to do the work. But I'm curious to know. Yeah. Where you see this headed.

Okay, so I think for AI, there are two parts. One part is where now we can build very simple chatbots that are trained specifically on certain accounting standards. So, for example, I have the international frs. I can feed it to the chatbot, and the chatbot can answer you according to the ifrs, and it doesn't go to the Internet to search for additional information, so it does not hallucinate so much. And the technique of prompt engineering will ensure that the robot will answer as accurately as possible. So this is something that's very easy to scale now. So you can have your Australian frs, and then this is something that you can make that can be shared with your team. The other place where I see AI can come in is that instead of the auditor doing the vouching manually, now actually you can use AI to help you to process the information from the invoices from the delivery orders, and now they can extract it out nicely into an Excel file for you. So I would say AI is like the brain, helps you to think, but you need rpa, which is the body that helps you to actually perform the work done. So how I integrate both together would be like, I have a folder, I will throw in my 10 different PDF files. I will get the RPA robot to read the 10 files. I will use the AI to pull out the data, invoice number, the date, so on and so forth. And then I'll get the RPA bot to paste it into a nice Excel file for you. So this will replace a lot of the copy and pasting of data for you.

Yeah, I love that. And I love, like, you've got a specific lens through which you look like you're looking into the finance department. And you've got these skills around RPA and AI and you're thinking about, and how does that get applied in the finance department and also for accountants and bookkeepers themselves, how do you approach or think about.

When and where to look for a solution? An RPA or an AI solution? Is it? Yeah, like you just thinking, well, what tasks are you currently doing? Or do you have a framework to try and think about how and where to apply.

Yeah, I actually have a process flowchart where you can prioritize based on different criteria. You can rank it for each criteria, whether the process is suitable or not. So the biggest thing I always, the first thing I ask normally is for the processes that you have identified, how long does it take for you to do it manually now? So if you only save 20 minutes in the whole month, this is not a process to be automated. I'm looking for things where it takes two days to perform a certain task and then that two days. With automation, it's just a matter of pressing a button one minute, the whole thing can be done. Yeah. So I look at the time savings first. Because if you're only going to save one hour, sometimes it is still cheaper to pay your accountant to do it manually than to go and invest in technology to automate it. So it's cost versus benefits exercise here. So once you know how much time you save and then we look at the volume and we look at how repetitive the task is. So if you have a thousand transactions to, to process, can the robot run and, and process the thousand transactions at night for you? If you can, that's good because then before you go to sleep, you can press the, the button. The robot will get the work done for you. Or when, before, when you go for lunch, you just press the button. The work is done when you come back. Yeah. So it's what, trying to see how repetitive things are, Are they volumous, are they rules based? Because if your process is very fixed, then it's very easy to build that automation and you can get a very, you can develop it quickly, you can achieve your ROI faster.

Yes. Yeah. And I, as I hear you explain, it's like that same framework could be applied in every department in business. I know you're looking very specifically for finance, but I think that same thinking. Yeah. Could work just as well in sales and HR and other departments in business. And I feel like it's, it's ripe for this opportunity. I think now that the technology is here and like you said, ChatGPT makes it a lot more accessible for people to look for these solutions. What type of use cases are you seeing? I know you kind of touched on a few, but what are some very obvious use cases?

Okay, so for example, you have the bank statement and maybe for every different transaction you need to classify the transactions into different expense and revenue categories. So in the past your accountant will have to search for different keywords, filter, and then you can categorize them in another column. But with AI, all you need to know is how to prompt ChatGPT for code. So the same process, this same bank statement process takes actually for one of my clients two days to actually go through each section to categorize them. Now it's just a matter of pressing three buttons, less than a minute. Everything is classified nicely for her. And the best thing is that when I did a project for her, when I checked back to her previous month's data, there was a lot of discrepancies. So I was a bit worried that, oh, maybe I didn't program it correctly, but actually the errors were due to the accountant who has some human errors when she did her work herself. So with automation, you're going to save time, you're going to reduce the human errors and then definitely you can focus on more value added at your work.

Oh, hey, I know you're tuned in right now, so I'll keep this super quick, but you like awesome YouTube videos, don't you? Well, there's not really that many YouTube channels that create great content about how to systemize your small business. So I want to make sure that you catch every video that I publish. So if you haven't already, firstly give this video the thumbs up and then go ahead and subscribe to my channel. You can even hit the bell notification to make sure that you're the first to know when I publish anything. So go ahead, do it right now. I'll wait a sec and then I'll transfer you back to the video.

Yeah. With accountants and bookkeepers, how do you see this changing their industry? I know there was a period there where at least we saw it in Australia where a lot of the back office tasks moved over to, you know, some different emerging economies like India and the Philippines. And some of those processes were happening, you know, back office offshore. And then you had the accountants that were really looking at, you know, trying to add more value and beyond just that because the industry was kind of getting a little bit commoditized there. And how do you see this now? I kind of feel like this is going to continue to shake up the industry as then some of those tasks that were previously offshore are now going to move to robots.

That is true. Certain things now you can have the full control. Also because let's say in the accounting audit firms, your data is supposed to be very confidential, so having those at your premise will be good instead of you offshoring the, the processes over.

Yeah, I guess where AI and RPA comes in Is the time is saved so that you can really come up with insights for your client. You can use AI actually to just throw in the balance sheet and P and L and then it can auto calculate the different ratios for you, your AR turnover, so on and so forth based on the ratios the AI can even assess for you. Are the ratios improving, are they getting worse? What are the possible reasons for the increase and decrease? And with this you're actually able to provide a more strategic advisory to your clients. And that's where the value add comes in for the industry.

You mentioned something around the security and data privacy, which is a lot of concern for business owners. I mean, yes, in the, the finance side of things, but really all throughout the organization. What are some solutions that you're looking at that either can be deployed locally or privately? Obviously the, a lot of the RPA stuff that's custom, very specific use case. When we think more about AI, have you started to explore what solutions might, might fit there?

Okay, so the basic level of security that you should have is when you use ChatGPT. You need to go into ChatGPT, into your settings, into data controls and turn off this option, which is improving the model for ChatGPT. You should turn it off. So whenever you talk in ChatGPT, your conversations and the data that you upload are not used to improve the model. So this gives you a first level of security on the data that you talk with ChatGPT. Now if you want the data to be really confidential, then you have to actually use the API key that's connected to ChatGPT and whatever data that goes through the API key, that data is secure, no one can access it. It's not used to train the model, so on, so forth. And ChatGPT and rather OpenAI and Microsoft actually have come up with a statement to say that that is the most secure way to protect your data. So I know how to build a simple robot that can connect to the API. And then in the long run you're only just paying for API tokens to run the automation. You don't have to pay for very expensive AI software and tokens now are so cheap. So if $100, you can really process a lot of information inside and use the GPT model to do many things for you.

Yeah, I love that.

What do you think are some ways that business owners should get started with this? I know you deal specifically with accountants and people in the finance space and there's probably some wins there in the finance department and you gave a good framework for really thinking about how and where you might prioritise what you start to look at. We always, I mean, our work is around process. So it's all about getting very clear what it is that you're actually doing. I think getting clear on where your time is going and then trying to prioritize from there and then thinking what AI solutions and robotic process automation solutions can fix that or help to automate some of that. When you're getting started with a business owner, what do you suggest to them or where they should start thinking first?

So I always ask them, what are your biggest pain points? And then they will list down maybe five. And then I will tell them I want to speak to the people on the ground, and then I'll ask the people on the ground what are their pain points? And the interesting thing is that normally what management thinks are the pain points may not always be the real pain points on the ground. So, for example, management may say that I've bought a very expensive software, but when actually I was watching what the staff did, they actually do a lot of very tedious things that are very manual that actually could have already been done in the software, but they're not using it properly. So there's a lot of inefficiencies within the team. There are a lot of things that are prepared. There are certain reports that are prepared. No one is reading them. There is a lot of data that's duplicated all over the place. So I think the first step is really to streamline the processes, who should be doing what processes, which processes should be cut, and then from there, once you have aligned the different job roles, then now you can see, okay, for this process, let's build the automation. And this is the time saver, the amount of time that we can save in the different processes. So once you do it properly, you can definitely save a lot of time in the organization.

Have you got any case studies of either clients maybe that you've worked with over a long period of time and what sort of wins can be gained?

So, for example, I have some people who have attended my training classes a few years back. So as the audit manager, she picked up the simple RPA skills. When she went back, she went to build her own robots that can send emails, prepare the confirmations, prepare different forms. Now she's promoted to head of technology in her company. So there's a way to transit into tech. There are people after the class, they have managed to build their own PDF extractor. They can really pour the information from PDF documents and they can save a lot of time because they can process a thousand documents every month very easily. Now many different use cases, I can talk another two hours of the different things that we have automated for companies and how much time they have saved. I think one thing is very important when I do the projects is the change management must be done properly. So when management is very involved in the projects, the projects will be much smoother because then everyone is really on board. The management has done to inform the change management to tell the staff that yes, we are using RPA because when staff hear that robots we're going to use robots, they will firstly panic because they fear that they're going to lose their jobs, they're going to be obsolete and stuff. But if change management is done properly to reassure staff that they are going to go home earlier, they're going to do less ot this is going to make their work life better. They're going to be doing more analytics, doing more higher value tasks. Their job roles will improve once everybody is on board. The implementation will be smoother. The digital transformation for your company will also be smoother.

Yeah, I think definitely people from team members perspective understanding that this is probably one of the best ways to be increasing your value at your organization because it just enables you to get done so much more and be so much more efficient and that then frees you up to work on those higher value tasks like that that really and any team member that is not using AI and rotobotic process automation versus those that aren't are and those that aren't like it's, it's going to be night and day difference. So those that do use the new technologies are going to shoot ahead and those that aren't will get left behind. So it's I think something that a lot of team members need to embrace. They don't really have much option here.

Yes. Yeah. The world is changing now. So I think every accounting and finance professional needs to learn how to read code. They are really how to like last time for us to write a VBA script to write macros that you need to have an IT colleague to really help you. But now with ChatGPT you are really empowered to do it yourself and being able to have your own macros running can really save you a lot of time. And that how I see finance departments in the future where at least half of the people will be able to build their own automation and then definitely you will see a lot of productivity gains from the.

And that's, you know we talked about a couple of like specific use cases. And you've got a lot. I don't know if there's any. And you talked about the, the lady moving up into that technology role from a, like a business perspective, it's so hard to quantify the wins. Like, it was great to talk with Shannon because she was saying, hey, here's one specific process that we automated. We saved a thousand hours. But when you're getting all of these wins throughout an organization and you're compounding and stacking those benefits, and then you're getting the team members to work on higher and higher value tasks, it's very hard to actually quantify what the benefit is. I don't know if you've come up with an easy way to kind of go, hey, we've worked with this client and literally X amount was saved, or if that's quite hard to calculate, it's doable.

So, for example, let's say for an audit firm, maybe the junior audit assistant gets about $3,000 a month. So if divided by 22 working days divided by eight hours, then you can quantify how much that rate is in hours. And then you just times the number of hours saved and you can get a very good calculation. So let me just quickly use my calculator. 3022. 8. Yeah. So you're saving about $17 an hour. Yeah. And then after that you can just multiply by the real time that is saved.

Yes. Yeah, yeah. I think it's exciting to see someone like yourself quite passionate, as I said, in the finance space and then starting to realize that this is happening in every industry, in every business around the world. I. I feel like this is the most amount of change I've seen in small business in such a short space in time. Like, if I look back to, you know, the Internet and Web 2.0 and, you know, Internet speeds, picking up like this for me is just, it's so significant. And I feel like where we're very early days, if we think a few years into the future, I don't know if you have a little bit of a crystal ball or any predictions on where you think this might go or trends that are worth getting on.

Okay. So I think one of the very good tools in the market that I think everyone should use is actually UiPath clipboard AI. So it's this. So what happens is that you can have an Excel file And then the UiPath clipboard AI can you just press one button to copy. You can go to your SAP system. When you press paste, the AI is able to know what value goes into what field and the robot will just paste, paste, paste, paste, paste all the data for you. Yeah. So I think the trend is you need to know what are the good tools in the market, what are the free tools that you can really utilize, and then how to use it in your processes. I think. Yeah. And there's a lot of very good tech in the market now, such as this, that you can really help you to increase productivity. Yeah. And then I guess you can join different chat groups and stuff, and then sometimes they will share a lot of very interesting AI use cases. And that's how I also learned about different applications to our work. But I'm always very mindful when there are new technologies that come up. It's always about cost versus the benefits because some AI software may be very expensive, and then again it's down to if you're not going to save a lot of time and then it's not justified on purchasing the software. So therefore I know quite a few of the free tools in the market that people can use. Yeah.

Yeah, I love that. Do you have any tips on avoiding chasing shiny objects? I feel like in the AI space, a new tool, a new feature, feels like it's released every day.

Yes.

So easy to go. Oh, that looks like it would be interesting. Oh, that looks like it'd be interesting. And I know in your space that's, that's what you do, you research to find those tools. Do you have any tips there?

I. Yeah, I think AI is not growing at a straight line. It's growing on an exponential curve. And I think the, the tools that have been there at the start have just become better and better. There is no advantage to building your own AI model now. That's just my take. Yeah. So.

What OpenAI has, using the GPT models, those are good enough to automate a lot of our accounting and finance processes. So I think it's just good to stick with the current tools that are available. There are a lot of new software in the market, but again, they want you to, you need to pay for those software. Sometimes I feel the cost is not justified. And then they have again another issue on the security and again whether your data is used to train their model. So.

I'm very happy to use ChatGPT because OpenAI and Microsoft have a partnership and you know that your data is safe based on their policy. So I think it's still good to stick to the ones that are commonly known. But I guess along the way people can listen to your podcast and then when there are new tools, you can always learn about it.

Yeah, perfect. Well, thank you so much for your time, Daryl. I'll pop a link through to skybots. Is that the best place if they people want to get in contact to reach you?

Yes, they can come to our website. Actually, on my website I also have another page where I have built various chatbots that are free for the accounting and finance industry. So, for example, I have a chatbot is trained only on the international frs. I got another one that's trained on the ethics standards in Singapore. So if you have certain Australian standards, you can always send me the PDF files. I can build it for free, I can put it on the website and people in Australia can use it and that should really help them.

Love it. All right, thank you so much, Daryl.

Thank you.