105. Will A.I. Destroy Fractional CFO’s?

Foreign.

Is AI going to replace fractional CFOs? That's what I'm talking about. On today's episode of the CFO Report, I'm having an interview with my friend Nicholas Bucher. For those of you that are in the finance industry, more than likely follow him on LinkedIn. He's got almost a million followers and today he's joined me for a really in depth conversation about AI and machine learning and ChatGPT. In today's episode, we're going to break down what the Differences are between AI and ChatGPT and all of those things. We're going to Talk about what AI and ChatGPT can do today and where the industry's going and what fractional CFOs can specifically do to either watch out for what's coming or to embrace that technology and use it as a competitive advantage. So without further ado, here's my conversation with my friend Nicholas Bucher. Nicholas, thank you so much for joining us today. How are you doing?

Good, thank you, Michael. Finally we meet again.

Yeah, we've been LinkedIn buddies now for a year and a half, maybe or two years, and we've had some really, really good conversations, you and I, we work in similar industry but very different things. You know me, I'm on the fractional CFO side. But you do something completely different in the finance world, Nicholas, if there's anyone listening that doesn't follow you on LinkedIn and doesn't know who you are, I don't know who those people would be. But for those people that may not know who you are, can you just tell us a little bit about what you do and what your background is in the finance industry?

I'm a finance guy, like everybody listening today. And I worked at PwC first, so I have this background of being an auditor and a consultant. And then I moved to the corporate side working in FPA and head of finance for Thales, which is a French company. And if you don't know about it, it's like Boeing but smaller. What really passionate me the last years is always share what I learn. And I discovered that on LinkedIn is the best place to do that professionally. Like if you want to learn about how you can cook, is not LinkedIn the best place, but for finance is really the best place. And I found that showing complex topics by using visuals is the best way to make people learn about it. And there is actually so many concepts that people still don't get because they didn't get the chance to have the exposure they didn't get the right mentors and I always use my channels of social media for this and lately because I really believe in it. I am focused now on showing how we can use AI in finance because I don't know about you but I really feel finance is really good at Excel. We are really good at using the ERPs and often in companies we are seen as the Excel or ERP gurus and I believe that for AI we can do the same and use AI to be a great department, bring a lot of value and then maybe also spread that around around the company and have more impact.

When it comes to generating CFO reports for my clients, my firm has two non negotiables. The process has to be accurate and it has to be efficient. That's why my firm only trusts G Accon. At its core, Giacon is a data connector between accounting platforms like QuickBooks and Xero and Google Sheets. But Gicon is so much more than just the data connector. Right out of the box you're going to get hundreds of templates for all the reports that CFOs and accountants need to run for their everyday workflows. Everything from balance sheets to PNLs to AR aging reports. Or you can use G Accon like my firm does to create our custom CFO reports every single month. My favorite part about gicon is that you can set it up to automatically create all of these reports. That means every month I'm getting fresh reports with fresh data ready to deliver to my clients. And because Gicon is automated, that has completely eliminated the need for my team to download data from QuickBooks, upload it into Google Sheets, manually apply formulas. All of those things that take up a lot of time and are prone to errors. In fact, G Accon has completely replaced all the other tools, services and workflows that we were using to generate CFO reports. Gicon is literally our one stop shop for making all of our CFO reports automatically every single month. If you're ready to start saving time and reducing errors, then you need to check out Gicon right now. Be sure to use the link in the description below to save 30%. That means you can get started with Gicon for as little as $28 a month. Join my firm and over 10,000 other firms and try G Accon today before we get too far into the weeds. I hear a lot of people throwing around terms like AI and machine learning and algorithms. I know you're not from the us. I'm basically from Alabama. I'm a country guy. I'M pretty simple. Can you explain the differ difference between algorithms and machine learning and AI? For people like me that don't know.

When you think about AI is basically the machine is thinking and doing stuff that our brain so deciding, analyzing, so replacing a bit of our intelligence. And also the machine is always continuously improving itself. So for example, a spreadsheet, Excel, it will always do the calculation between two cells and will not be creative of thinking, oh, actually there is a better way, a better formula. No, it will always work like it's programmed. Then when you go down to AI, you have different types of AI. For example, machine learning, it's based on the data you put in the machine learning, it will keep continuing learn and provide more input for the users based on the data. So if you keep showing the AI pictures of dogs and you keep also training, saying this is a dog, this is a dog, this is a dog. At one point it will build algorithm that will recognize pictures of dogs. What we are using right now, all of us is the generative AI, which use nlp. So natural Language processing, it's algorithm that understand how you and me, we speak and will talk to the computer, say oh, Michael wanted to know about the population of Germany. And then it will look into the model, will do some mathematical calculation to check where it is, what is the most probable answer, and then give us back the answer in our own words. And that's what actually was the big change last year, when people that never knew how to code, never knew how to talk to AI, were able to just chat with models that have a huge knowledge. And instead of going in Python and coding, you just ask a question and you get a reply in your own words. And that's why since one year, AI is everywhere, because it just got more accessible for all of us.

Would it be fair to say that algorithms are equations? AI is taking algorithms and applying them to make itself smarter. It's learning how to take algorithms and apply them in different ways. And now we have things like ChatGPT, which has got algorithms that are able to decipher what we're saying. When we use normal human language, it's applying algorithms that are converting what we're saying so that the AI can then go figure out which other algorithms it needs to apply to get the best answers. And then along the way it's able to apply other algorithms to make it smarter and to give better responses. Is that a fair breakdown?

Those algorithms, they have a name. It's the in GPT, it's the T is a transformer. So how you get an input, you put the algorithm in the middle and you get an output. And those transformers, they got also, a lot of them open source. So it's like if you and me, we work on this recognition of a dog and we figure out what is the algorithm to recognize dogs, instead of keeping it for ourselves, we'll put that in the public. And then other people who have maybe training dogs, and they will put their algorithm plus our algorithm and make a better product. And that's why recently it's going so fast is a lot of people are sharing those algorithms for free. And instead of having to always restart from the beginning and refigure out. Okay, how do I go to from A to Z? You just do an addition of all of these transformers together and you improve all the time. And we just get exponential growth in terms of the progress. Because now you also make AI write its own code so it knows how to improve itself. The only limit is two factors, is the data that you can provide. You and me, like, as individual, we don't have much data. But recently, like Reddit was selling to Google all of their data on all of their chats, and Google is using this to train its model. And the second limitation is the power. So how much processing power you have. And that's why the Nvidia shares are so high, because there is not enough processing power for all of the demand.

I'm an Apple guy, right? And so a lot of times I'll ask Siri, hey, what's the weather in Washington? And Siri says, the flight to Boston doesn't leave for two hours. I mean, it's. There's such a massive disconnect in the things that I'm saying. And so one of the questions that I have for you is when we're thinking about AI, particularly in terms of the financial industry, it's sometimes hard for me to wrap my head around the fact that as an industry, we need to be embracing AI, because I see how far away Siri is, and honestly, some are better, but there still seems like there's a huge gap. And so when I think of how do we apply that in an industry where accuracy matters? Where is AI in the world of finance today? And how important is it, would you.

Say before ChatGPT arrived? And when I say ChatGPT is really the chat interface that allow us to speak to it without coding. Before, you still had banks, insurances, funds that were using AI with machine learning to process a lot of data. And because our brain and Excel is not enough to process all of this data, they use machine learning to recognize patterns, to make sometimes investment decisions, or also to identify fraud. So that was already there before and is not new. You have also a second component which is recognizing invoices and digitalizing all of the P2P process. And you had the step until the computer recognizing all of the fields on the invoice, invoice number, dates. But where the computer could not go further was to say, where should they put that in our erp? And there the human was there to do rule based matching. All of this was there until one year ago. Now, since one year, you have the huge progress on nlp. So the language, meaning that also when we talk and you can check like all of the voice recognition works now perfectly. If you have videos and you get subtitles. It's really, really rare. Even for me, who I have a really bad English accent. As a French guy, it recognizes. So I think Siri is just outdated because everywhere where you go, it recognizes what you say. So now on this, the language is really good. And because of this, today we talk together with words. We communicated before you and me, with emails. You might send to your client a report on the findings of the month. You might write procedures on how to do the reconciliation of your bank. You might analyze legal documents. Everything that I just described is with words and not with figures. So we spend, even though we are finance guys, we spend most of our time working with words. Because right now generative AI is really, really good at understanding and generating words or changing them to have more impact. There is a lot of use cases. So what I teach to people is first, how can you prompt the right way that you get an output that you can use? So there are some prompt techniques like just giving the context, being specific and instruction. I call it the csi. And then if you bring the FBI team, you tell which format you want, you give a blueprint. And then you can also ask AI to have an identity. Then you have CSI and FBI together and you can use this. And this can help you create procedure. So you can ask AI to create a procedure with you. It's just like having a guy or girl who wrote millions of procedures and you just ask, oh, write me that. And on top of knowing how to write it, they will write it in one second. Because it's really fast at writing. If you know how to ask, you can create procedures, then you can create templates, you can create checklists, you can solve problems with AI. But for this you need to have a good brain because you need to think in frameworks. What type of issues do one of your clients had recently and that you had to dig with your, with your team?

We were recently working with a client. Their gross profit margins are significantly too tight for where they need to be. So their overhead is really lean, revenue strong, but their gross profit margins are too tight. So that's something that we're helping a client battle with right now.

So their cogs were too high, Cost of goods sold were too high?

Correct.

Then what you will do is you will go into ChatGPT and you will start a discussion where you say, okay, this is the context. I have a client. So don't put any confidential information. Because using ChatGPT is like using Facebook is like using Google. You don't own the data that you put inside. But you can start being generalist. If your client is a SaaS company, you can explain, okay, I have this SaaS company, their problem, their cogs is too high. I want you to help me to draft an action plan and it will start Giving you like 10 to 20 points that are quite general. And that's where a lot of people will stop because they will think I didn't learn anything there. Of course I have to reduce cost. Of course I have to increase my price. And that's where people stop. And they don't understand that you have limitation in the output of ChatGPT. So ChatGPT wants to please us, but cannot please us in one go. So instead you have to go like step by step. And in the 10 or 20 actions that you identified with ChatGPT, then you are going to pick up one. And let's imagine that you are talking about looking at your suppliers and benchmarking the price of your raw materials or price of subcontractors. You pick that and you say, okay, let's go down. So you go one level down. Let's analyze that together. What can we do on this? And then it will give you 20 other actions that you could do just on this topic. And then inside this there might be one that are negotiate with another company to partner to have a bigger purchasing price. And then you're like, oh, this is interesting because I have actually a friend who buys for this company and we could do that. So then you go on this one, you ask again, let's drill down. And now you have a real action plan that you can use. You can even use it after and say, draft me an email that I can send to my partner company to start negotiating on this. And you went from having a Big problem to having a concrete action that you can use. And this just needs that you drill down three to four to five levels and that your mind is also problem solving oriented to go to it. Because you know how to drill down and you know how to apply it after in your work.

I'll be honest, I'm the guy that stops after the first one. So I'm like, okay, I pay for the upgraded ChatGPT, right? Because I want four. If four is the latest one, I'll pay 20 bucks a month for that. I would go in and test it out and say, you know, hey, here's the problem. They give me a bunch of generic solutions and I'm like, well, this is trash because it's generic. So what you're saying is, the trick is taking what we know about the context of our client, let's take a look at that list of 10, 15, 20 generic recommendations. Find a small few that in the context of that business, might make a lot of sense. One by one, let's drill down another layer. Look for one or two or three of those generic answers until we get to the point where there's something actionable we could document an action plan from. And then we just kind of take the three, four or five of those and now we can put it into a report or an email or some kind of a tracker. I've not gone that far, so I'm immediately interested. Do you find, Nicholas, that it understands the context of industries? The suggestions they make are typically in alignment. So, you know, we had mentioned SaaS companies, but maybe it's an electrician or a plumber or law firm. Does it understand industry context if you.

Provide it at the beginning, if we talk about SaaS, it will start answering for a SaaS company and maybe that at one point in the discussion you derive on negotiating on a supplier. It's not that it forgot, but it focused the discussion on the supplier negotiation. And then it will come back to be more generic. And what you need to understand is in the discussion, there is a memory it understands and remembers the context. If you go too far in the discussion, it's optimizing itself to keep only what is necessary. If you feel that it starts to forget, you need to remind, oh, be more specific about SaaS company. Each time you see that it doesn't go in the right direction, you just correct it. And it's so fast, by just changing one word, that gives the context that you always can bring back ChatGPT to the right track. And that's also again, the trick, because people will think, oh, it forgot and they will stop. Never accept the first answer as the right one. Always try again, do something else. And often you will get to your results.

That's really helpful. And let's shift the discussion a little bit from recommendations. In your experience, how good is it at analyzing a set of financial reports? Again, given industry context? So if we took, you know, anonymous set of income statement, balance sheet, statement of cash flows, this is a SaaS company, is it able to analyze the financial statements and provide meaningful insights?

Now we went to the right question, which is what about figures? Because of the nature of the model, the model always wants to generate the next word or basically the next letter or next word. So if you say, I am now in Berlin, it's February, the weather is. And the AI will generate the next sentence based on the model that it has of a lot of times that they read the sentence, okay, Berlin, February, most probably the weather is cold. When you think about the financial statements, it's not because you see that The GNA are $533,000 that the RD the line under will be $534,000 or portion of this. It's not like this. We need to get out to the idea that just by plugging the figures, it will do the work of an analyst. Because an analyst will go to Excel, will do calculations. But what it can do, it can help us understand what are the right KPIs to calculate, what are the right methodology to analyze some figures. For example, if I am in SaaS, I will do cohort analysis, which I will not do. If I have a manufacturing company, I will look at my standard cost and maybe my scrap rate. I am a SaaS company, how can I analyze my revenue? Then it will start bringing you to the cohort analysis and explain you the cohort analysis. But then if you want help on the work for ChatGPT, you can already upload your data because you are not allowed. And even if you are allowed, because some companies have their own GPT, there is so much data to process and it's not good at calculating because again, it's really good at thinking what is the next word. So because of this, you are going to ask the model to tell you how to do the analysis. If you give enough context of saying, okay, those are my columns in column A, I have this B, C, D, it will exactly give you the formula to apply or how you can visualize a cohort analysis with a heat map. You will start from not knowing how to do it to in five to 10 minutes, having the blueprint on how to do it and in your environment, on your file, scalable and the second step. And that's bad news for all of the finance people. But I have the good news behind. We will have to start using Python because this is the door for automation and financial analysis at scale. The good news is we will not have to program Python because chatgpt will do it for us. If we tell chatgpt, can you code how? I do a heatmap in Python and my data looks like this. So giving the context again, it will create exactly the code that will generate the heatmap in Python and then you just copy and paste the code in your protected environment, you execute the code and you get the output.

How accurate is it today at generating useful code that's copy and pasteable into the Python environment? Is it good at that? Is it great at that? Is it okay, but it needs work.

That's a good question because also you as a user needs kind of reliability on this. Combining files, creating visualizations. It's quite good at it. If you start to need to have the context and all of the data, you will need a bit of iterations and tests, maybe. One story about myself, I never learned Python. I never learned programming because I'm so deep into the topic and saw the potential. I used it two months ago when my accountant told me, those 80 files that you have with 100 lines by file, I will need one week to upload that in our accounting system and process it because we need to change all of the lines. Sometimes they are negative, sometimes positive. It doesn't work with our accounting system, so I need to change it. So I really thought, okay, I'm sure in Excel I could do that in half a day, but then I will have to do that half a day every quarter. And then I thought, this is a really good case for ChatGPT and Python together. So I started first by asking ChatGPT how do I combine 80 files? And that was like three lines of code. I just applied the code in my Google environment and BIM the 80 files became one. Then I went into ChatGPT and explained, oh, I want to add a column to give the account number, I want to add a column to say if it's debit or credit. And then if you see PayPal, like a line with PayPal, just remove it. So that was starting to be a bit hard because it's a lot of data cleaning and with just two or three iterations, of course, test it after to see if it worked, I explained, oh, I forgot you to tell that this code needs to be like this. So that I forgot that we need to remove a line. I tried again and the file became from 80 files, it became just one file. I call my accountant, I say try. And then on the phone he tells me in one second that the upload worked and everything is green. And so you can imagine 40 hours by $100 that my accountant cost me. And I just needed one hour to automate that.

The important thing is starting to learn if somebody isn't using ChatGPT and those things. Now, it doesn't sound like based on what it can and can't do with data, what it can and can't do with contextualizing a client, all those things. It's not so much that you're going to lose your job in the near future to these programs, but it is definitely something that can make your life a lot easier or faster or more efficient. It sounds to me like what you're saying, Nicholas is the job today is to just get in there and start using it and collaborating with others that are using it to figure out where are those limitations, where are the benefits and how can you apply those in the context of your firm? Is that an accurate description of where we're at today?

Understanding what can it do? So what we did today, the discussion because you need to understand the limitations and like a spreadsheet, a spreadsheet will not make a video for you because AI sounds like magic. Everybody thinks that AI can do everything. So you need to take a step back and understand how the model works, what it can do, what it cannot do. And once you know this, then you will need to work with other tools to make it that from the start of your program to the end, you can still incorporate AI in some of the part of the process to make your whole process faster and better. The more we will move towards evolutions, the more AI will absorb part of the process because they will start connect with other interface and the algorithm will start to understand better better the context. But we are not there yet. If you as a human are the maestro, then you can use AI really well and make your process smoother and faster and deliver more value. But if you ignore that it exists, might be that your competition is doing it and selling services at a lower price to your clients or delivering more for the same price. This is more I would say on how much value can you add in which time you need is more this equation than all of us Losing our jobs.

Now that's where we're at today. I know that earlier today maybe it was Microsoft came out with pretty big announcement around AI and finance. Could you talk a little bit about what was that announcement and where do you see GPT in AI in a year from now or two years from now? Where are we headed in the short term and the medium term?

What Microsoft is doing is keeping everything in their own environment and to start making all of the modules talk to each other. And that's the best use case to use AI because when AI can start understanding the data, it will learn how to talk to all of the apps. And the announcement of today was Copilot for finance and it's basically Microsoft connecting AI plus Excel plus Power Automates plus Microsoft Dynamics or they say also SAP, but I guess Microsoft Dynamics first they have also a tool in Power Automate that recognize invoices and can put that in a flat table. So they connect all of these tools and will make our work. So for example, a bank reconciliation faster because all of the processing, all of the data preparation and the matching, they will do the preparation and the big work. And then when it comes about having the context, talking to the client or talking to the supply chain, then we can take over. Microsoft can do that because they have a finance system. So Microsoft Dynamics, they have invoice recognition system, they have AI, they have Excel, and they can connect all of these tools together. And that's where I feel the AI is going to go is when you can have interface between AI and all of the other tools that you are using. If you are using QuickBooks today and then you have your treasury tool plugged into QuickBooks, you have your CRM tools. Well, either QuickBooks will come up with an AI layer that will allow you to ask the questions and the answer will come back from an aggregation of all of these systems. Or there will be an external solution that will come say, okay, I know the API of QuickBooks, I know the API of the treasury system, the CRM, and I will basically plug my AI on these APIs to give you answers that usually takes you weeks to get or maybe you don't get today because you don't have the time to process those questions.

Sounds like with that there are jobs potentially at stake because there are jobs today that do the reconciliations and the matching and those kinds of things. How good is that today? Have you had a chance to look at the Copilot for Finance?

I just saw a demo. I'm sure if we take a company that is already today, not really digital. There are so many low hanging fruits, so they will be the first to benefit from it. If you are already so far in your automation and process, you might not benefit straight away from it. But what is good right now, if you didn't start your digital transformation, you can jump straight away on the last solution.

Probably if you haven't started a digital transformation yet, you don't know what Bard is or Copilot, you don't even know what that stuff is. So I don't know that there's anybody that was sitting around like, you know what Bill, Now's the time. Now's the time to stop, you know, hand printing all of our invoices and writing checks. But who knows, maybe Bill's out there. I have no idea.

I didn't see the solution. So I'm really excited. Usually like Copilot for Excel, it's at, I would say the surface level, it looks good, but once you come to the, you know, like the details in the devil is in the details. When you come at this detail, you still need humans to come up and.

Finish the work today anyway, let's shift gears if we could. One of the struggles that fractional CFOs as an industry has is social media. Figuring out how to use social media, you've done as good a job as anybody I know in real life, like where I actually know the people. What is the role of AI when it comes to. Let's just use LinkedIn. That's like both of our platforms. I have some hot takes on AI when it comes to creating content that I think are a little bit against the grain in the industry. But you and I have actually never talked about this. So I'm curious, like what is your take on how are you using AI in content generation for LinkedIn? Let's just start with that question. How do you use it today?

Maybe a little secret. I think when it really started chatgpt, I did some of the posts with this that performed so well. Like I did the CFO checklist. It was just a text with just output from ChatGPT, but it was not like one click. Like I think I processed it. I clicked 20 times until I got 20 good items of what should be on the checklist of the cfo. That's why I say to everybody, like everything you are going to use with this tool becomes yours. If it becomes yours, it's your own responsibility to own the content. If I am using Excel to do a financial statement and three ways model, I cannot Blame Excel to wrongly calculate or wrongly match two sales. It's my work. So either it looks good and it's working well, everybody can clap, or if it works, not good, like it's my fault and not Excel. And for me, ChatGPT is the same. I am not English speaker. I've learned a lot about copywriting, but if I have a really good idea in mind, it's most probably better that I help myself with ChatGPT than write everything by myself and also for my audience. Because if my idea gets through much More easily using ChatGPT than just my own words, man, it's only a win win. All of the work I did gets understood better and by more people. That's my take on this. We are all of us now so aware that everything you can do that you can see if everything was done only with ChatGPT and nobody did an effort, you can see also some of the formulas replicated by everybody. Now the trick is to write with your handwriting on a board something that could come out of ChatGPT to make it feel more human again. Personal branding, you need to be always there. People forget you. And if you can make your work and your post faster, but with the same quality, then why not using the right tool for this when it comes.

To the personal branding, and I think that's where LinkedIn is amazing. Yes, it's a professional social network, but it's really the personal branding side of it, I think that builds engaged and loyal audiences. Chad, GPT is kind of the opposite of that to me. It's the opposite of the personal, authentic, vulnerable kind of thing. So how have you found that you can marry those two, like, in a lot of ways, diametrically opposed things into something that your audience clearly resonates with.

Before ChatGPT, there was already the. The bromance. Do you know this? When people were just doing, like jumping every line to say a story, it was just like copywriting techniques that people apply. And all of the posts and on Twitter, the threads, everything looked the same, like top 10 books. If the books are good, it's good to share that and it's not a problem. But then you don't make a connection with people. And that's why I always had that. I put my face everywhere because I am still a human behind. I put my name and my colors, and when people will see my colors, they will stop. And people always tell me that, like, I know it's you, like it's your design. And other people started to copy that, but it's okay. You copy things that work. That's also why every time there is somebody who needs help, like some of my students, I jump really quickly on the phone with them because I want to make connection. Every time I'm connecting with people that are active on LinkedIn I also get on the phone like we are doing right now. We can post our words like we want at the end like what we are experiencing. You and me. This discussion, this cannot be replaced. I can spend much more time investing in this than just post. Trying to craft a text post which really sounds like me. I prefer to provide as much value as I can in the best digestible way. So with a nice graphic and then create real human connections in one to one or in group. Today everybody connecting with us, they are in their car, they are listening that with their ears. There's nobody doing the summary for them, I hope. And you and me, we are having a discussion and we know each other. We chat once in a while. We like I know about your family, you know about my family. It's part of the personal brand as well, is being there as a human in front of people. And I could also like find a way to record myself as an avatar that's not there yet. We'll not never have my accent where.

I found AI and ChatGPT to be helpful. Great ideation. Hey, I want to write a post for other fractional CFOs about the top 10 ways they can do this. I'll go in there and I've already got some ideas, but I might scan that ChatGPT comes up with and I'm like oh, number 18. I never thought of that, but that's actually a very legitimate thing and I'll pull from that. And so it's not about replacing. You don't use it to try to replace you in those things that make you. You use it as a supplemental tool for getting additional ideas or different approaches. Is that an accurate summation? Nicholas?

You learn a second time. And so when you need to prepare a post like we are not a master that knows everything that I've read all of the finance books. We use our experience and our understanding of things, but when we want to explain to other people, we need to do the research, we need to work and make sure that what we say is not wrong and that there is maybe not another way of doing it. A part of this is using tools, Google books, other people that posted on it, other blogs and chatgpt to learn about the topics and at the end using tools to formulate all of that in the best way possible.

I'll tell you one of my favorite things about LinkedIn that wouldn't be possible with just the application of ChatGPT is having the opportunity to meet people like you in the industry. We've known each other a year and a half, maybe two years. I don't know what it is now, but I've got a chance to meet you and learn. Just a great guy you are. And like you said, we've learned about each other's families. But my point is like, we can't circumvent the human side of social media. Like the social side of social media. You're listening to this and you're like trying to find the hack. Like, I just want ChatGPT to write my post. I think it's going to fall flat for you.

Like I said at the beginning, my post when I was leveraging a lot, but I was still reviewing it and owning it, it worked really well. Now people are expecting personal story. They want actually not everything that you can do in one second. They want to see that you did the effort and if it looks like you didn't do the effort, people will just discard it.

I need to learn more about AI and I didn't know coming into this conversation where it was going to go. I was actually talking to somebody in the hall earlier. I was like, I don't like AI and I don't think there's really that many use cases for it. How can I use it to get better ideas, to work with the client that's having the cost of goods issues? Don't you have a program for people like we can come in and learn about it and kind of collaborate as a community. Can you tell us a little bit about what that is? Nicholas?

I created actually this program called the AI Finance Club. It's changing so much AI. I'm also learning myself every day so much. And I want to share that with fractional CFOs and CFOs. We have already a lot of fractional CFOs and CFCs in the, in the program. With this program, what is good is we kind of mutualize the efforts. So I do the like 99% of the work with some experts as well. Once a month we have a masterclass. Last week we learned how to use Azure in finance. So like this we discovered how a small company can use Azure and use it for their P2P process. I organize all of this. I have the experts. Today I send an email about everything that happened in February. So I digested all of the news. I have like 15 newsletters that I follow and I follow like 100 people and I just sent five really important points to the members like this. They don't have to follow, they don't have to spend time to research. They just get everything from me. Also inside the membership you have all of my courses on AI. My goal with this is if somebody knows that AI is important for them for their work, but they don't have the time to invest and learn and they don't have the time to follow all of these people or to go in like 10 different courses or watching YouTube videos every night about the latest ChatGPT changes. That's why I did this program. What has surprised me is the quality of the people that are inside. We have a lot of fractional CFOs, like I said, mostly Americans, but some Europeans. And we have a guy who is the head of transformation at PayPal. We have also the ex CFO of Oracle that is in the group. And each time I find AI and finance experts, I bring him in as well to make sure that everybody gets exposed to this person.

Today I learned what AI means, so I'm probably an expert now. So if you ever need me to come in and explain to your audience what AI is, I'm available to hire. No, in all seriousness, if somebody wants to learn more about the AI Finance Club, which I'm one of them, how can we sign up or learn more about it? Nicholas.

So I will give you the link, Michael. So like this, you can put it in your show notes. If not, just go on my page on LinkedIn you will see straight away the AI Finance Club logo. Or on Google if you search AI Finance Club, you'll find it.

Final question before we go. Is there a world where Google Sheets replaces Excel in the finance industry?

Well, now that Microsoft is winning all of the battles on AI, I'm not sure. So Google is now a bit behind because of this.

I would say they're not pulling up the rear though. Apple's firmly in last place, so don't worry. Google Sheets, thank you so much for being here. I know it's late your time. The thing that I always think about when somebody brings you up or when I see your post. A kind servant hearted person you are and what a blessing you are to the industry. So thank you for being one of those people. Thank you for being you and it's just been an honor to chat with you today and I can't thank you enough.

Thank you, Michael. And thank you for doing that. Also for all of your audience to bring people to share knowledge. And also, I know that you are being an inspiration for a lot of people that want to become a successful fractional cfo. And I know how you are. You have been doing the work and now like showing to everybody how to do it. So I really appreciate that I had the chance to be on your show.

Awesome. Talk to you later, brother. All right, my friends, I hope you found this episode helpful. If you did, it would mean the world to me if you would leave a five star review on Apple Podcasts, Google Podcasts, Spotify, wherever you're listening to this episode. In the meantime, I can't wait to see you back right here next week. I'll see you then.