Video: Better Together presents: Data Storytelling with AI | Duration: 4208s | Summary: Better Together presents: Data Storytelling with AI | Chapters: Welcome and Introductions (23.235s), Session Overview (104.895s), Session Overview (188.995s), Data Storytelling Fundamentals (391.235s), Human Intelligence Side (696.355s), Finding the Healthy Middle (1104.935s), AI-Powered Financial Analysis (1815.965s), Data Storytelling Fundamentals (2648.155s), Data Story Framework (2856.9950000000003s), Data Story Creation (2968.865s), Refining the Chart (3149.365s), Real-World Impact (3426.415s), Practice and Next Steps (3547.515s), Resources and Community (3760.025s), Wrap-Up and Next Steps (3819.585s)
Transcript for "Better Together presents: Data Storytelling with AI": Hey, everybody. Welcome to our AI for financial storytelling webinar. We're so excited to have you here today. My name is Artie Martinez, and I am an education and engagement leader for Intuit accounting. And I am excited to introduce you to our host and our instructors today. We have Alicia Eastfold with us, and Alicia is a CPA who spent her career helping nonprofits navigate their accounting technology and has recently served on Intuit's partner council. She recently stepped out on her own to found the human accountant where she teaches accountants how to use AI to do more than close the books and how to find the story in the numbers and help run the business. Her first test of that idea is a place where the stakes are real, her dad's forty year old machine shop. And today, she's bringing you onto the shop floor. And, also, we've got Nio. Making accounting a little less taxing is Nio Carter Gray's goal as owner and founder of First Step Accounting. A self proclaimed techie, Nio decided a virtual accounting practice was the best way to travel and still helps small business owners all across The US manage their financial systems. Nio is an enrolled agent, former Intuit accountant council member, and is currently the chair of the Maryland Board of Individual Tax Preparers. Okay. So we're so excited to have you all here, and we wanna talk a little bit about why today matters. Before we jump in, I want to acknowledge that everyone is coming in today's session from a different place. Some of you are already using AI regularly, and some are just getting started, and some are simply curious about what's possible. Wherever you are, you're in the right place. We're going to cover a lot today. Some of it may be familiar and some may feel like drinking from a fire hose and that's totally okay. Don't worry about capturing every detail. You'll have access to all of these slides, the workbook, and the recording afterwards. So my encouragement is just to be present, stay curious, and enjoy how AI is being used in real world situations to bring more value to clients, and, hopefully, you'll walk away with something you're excited to try. We've also put together a workbook with examples, tools, and activities to help you put what you learned today into practice. You'll find it in the resource section at the top right hand side of your screen where you can download both the workbook and today's presentation slides. So don't feel like you need to capture everything today. Just be present, take it all in, and know that you'll have these resources to come back to when you're ready to give it a try. Lastly, a little bit of housekeeping. We will have polls for CPE, so go ahead and check the right hand side of your screen. You'll see a little red dot above the poll, and you'll be able to click that and answer your polls. The workbook is also going to be located there under the resources section and the slide deck that you'll be able to download. And with that, I'm gonna hand it off to Naya. Thank you, Ari. Hello, everyone. Super excited to be here and to learn from Alicia today. Before we get started, I know that CPE is super important. So we're gonna start off with a polling question, and this is to get your wiffle wet on how to get that CPE and to help us understand where you're starting at today. So which part of data storytelling with AI are you most interested to find out more about today? Is it strengthening your client discovery processes? Is it deepening your analysis? Or is it shaping the financial story? I can tell you I am super excited for myself to learn more about deepening your analysis. We're using AI in my practice, but now I am looking forward to seeing how Alicia, who has taken over that business to modernize it, is going to share with us some insights on doing that as well. So which one is working for you? I'm gonna close this polling question. Where we're going today? Just so you have a framework and knowing exactly where we are and where we're going. We're gonna start with discovery, and this is how we're gonna play like detectives with our clients. So we're gonna figure out what data story needs to be told the most. Then we're gonna do some analysis work where you're gonna see firsthand how you can use AI to analyze your client's financial information. And then lastly, we're gonna dive into the storytelling. And we'll watch as AI helps turn that analysis and discovery into a powerful data story. So here is where I hand it over to Alicia. Ready to go? Thank you so much. I cannot tell you how excited I am to be here. Now you've probably heard about all the amazing things that AI can do for accountants. Maybe you've seen some pretty shiny demos, but you didn't really understand how all of the AI worked, or maybe you have heard people talking about how it's making their work go so so so fast. Now I gotta tell you, today, I am not gonna give you a shiny demo. I'm gonna show you a gritty one up close and real from my dad's machine shop, and you're gonna see how AI gets it wrong. You're gonna see a lot of back and forth, and you're gonna see times that it slowed me down instead of speeding me up. But you are also going to see how it changed the way my dad's business runs, and I'm hoping you see how we can start imagining a bigger definition of what it means to be an accountant, a human accountant. Now I gotta tell you, doing this has been some of the most profound work I've ever done. I didn't get... I didn't check my brain out at the door, and I didn't just hand it to AI. I didn't become less of an accountant. I got to use what I've learned as an accountant to help shape a business. And you might be somebody who serves a lot of clients. Maybe you're a part of an accounting firm, or you might be the one finance person inside of a business or just a business leader. Doesn't matter. Whatever your role, there is something here that you can use today. And there's a way for you to stay involved, and I'm gonna talk about it at the end. So let's dive in. We are made to tell good stories. Now I love a juicy one. Pride and Prejudice is my personal favorite because it's full of drama and suspense. And Darcy has lots of pride, and Elizabeth has lots of prejudice. But all of these stories, they have conflict and problems that need to be solved. And the reality is there are really good stories, and they are buried in the numbers and in the pain points of the businesses that we're serving. They're hiding, and they are waiting for us to find them, like little gold nuggets. And our businesses, they've always needed more from us than just closing their books. The problem is we haven't had a lot of time because we're so busy trying to get all those numbers in order, and it takes a lot of time. So often because of that, they often get generic reports. But the reality is these businesses, they need their story told. Now I've got some sad news. This webinar, it is called AI for data storytelling, and I gotta tell you something. AI cannot tell that story. It can't read the room. It can't be accountable for the mistakes, and it also can't decide what part matters. You are the right person to tell this story, and you are a better storyteller than AI. But AI can support you, and today, I'm gonna show you how. Now you've heard a little bit about me from Artie, and here's the part that we didn't get into yet, which is why. So many are estimating that AI is aiming to automate about 95% of our regular accounting tasks. And in my career, I've got to have a front row seat to seeing where that's all going. And when I started to see it, it really sent me back to my why. Why did I become an accountant? Now it wasn't so that I could get through the work 95 faster and then serve, like, a 100 businesses. I became an accountant because I wanted to use numbers to make a change in the business, not just report them. Now here's the thing that I've always felt. Financial reports, they are the lagging indicator in a business. If you wanna change what caused the financials, you gotta paddle up the stream. You gotta figure out why the numbers are the way they are, and you need to help the business do something about it. And that's what I wanted to do. I believe AI is gonna help us do that better than we have ever been able to do before. Now I was a teacher a long time, and before I was a CPA and an accountant. So I started my own company, The Human Accountant, because I wanna teach other accounts how to do this too. And I also started experimenting a little bit closer to home, and that's with my dad's 40 year old machine shop. He reached out to me, and it's the business with the most skin in the game for me I have ever worked in. And it has been some of the most rewarding work of my life. So let's jump into how I was doing this. Before I could answer anything for my dad, I had to figure out this question that needed answering, and a lot of that happened through discovery. So to tell the story in a business's data, you gotta figure out what problem they need solved. It's not always handed to us. We have to play detective. Now clients and businesses, they give us clues, especially at the beginning. It's usually the reason that they reached out in the first place to us. Now if you listen through many conversations and find what's keeping them up at night, usually, they're saying, I couldn't sleep. That's usually the right place to start. Everybody is on the edge of their seat for the story that's found in those things. And the clues come out in little sentences here and there. The other clue is the financials. Numbers, especially reconciled and orderly ones, they don't lie. Alright. So let's talk about this shop a little bit. So forty years in, it's gotten really complex over the years. There are now robots. There are lasers. There's 20 to 40 jobs at a time, and they are each going through 20 steps before it ever reaches a customer. And my dad, he reached out because he wanted to get back to inventing, and he is stuck in the slog of operations. And reality is sales were softening a bit. Nobody could see clearly where to focus with all this complexity. And the team had a big question. Are the small jobs slowing down our sales? The worry was that they took a lot of work and didn't pay off. And everyone who has run a company knows this challenge, but the shop, they love their small customers. They just needed a way for it to be sustainable. And the answer to that question wasn't what any of us thought. I needed to be a detective. So I wanna show you what AI did to help me be one. So I'm gonna go back and forth between what AI did and what I did. So I am so excited about the ways I used AI. First of all, it was my note taker and my stenographer. I recorded meetings with all of the staff, voice memos in the car, with my dad on the way to a pickleball game or even sitting next to a machine while the guys are running parts, and it recorded it all. I had days and days of conversation. But then it was my memory keeper. It held hours and hours of conversations in one spot, and then it was my pattern finder. Overall, this conversation, it mapped 41 different pain points that I needed to address. It also even caught me. I kept sub... Suggesting something that I wanted to build, and AI, after I would ask it, would say, your dad has mentioned this other thing, like, four or five times. Why aren't you listening? It was also my thought partner. And especially being a solo owner now, this was super helpful. It helped me prepare my questions for every person that I talked to, and it coached me through manufacturing, an industry I hadn't worked in, it was pretty amazing. Alright. Now I wanna talk about the human side. What did I do? Because this is just as important. The human intelligence side is the side that needed me, and I had to be the one to build trust. Now this shop, they run machines. They are not in online meetings. They're not used to being recorded. And I realized that quickly, and I didn't want surprises, and I wanted trust and buy in. So my phone was on the table. I was discussing with every person why I was recording and where the information was going. And if they said something that they didn't like, we could delete it. I wrote out some of these guidelines for this, and they're in the workbook. It could help you if you're in this situation. Now accountants, we spend our situation. Now accountants, we spend our careers living in internal controls often to protect things like cash, and that's a big part of our trust. And I want us to think about data and protecting our business' data. With AI, that is how we build trust is protecting their data. So before anything goes into AI, I make sure that the tool I'm using is not training on private information and that that information cannot leak. And in practice, that means a paid business account or plan with the training setting turned off. And the reality is AI is now in a lot of our tools, so you need to read the fine print on how any tool handles your data. Next is I gotta... I had to read the room. I saw their faces. I could hear the edge in the voice. I could see them wincing. That's something that a transcript doesn't catch. I had to read the room and have empathy. And then, ultimately, I had to be the one to help decide what matters. I had an idea of what I was gonna first build. It was a cash forecast. But I got a phone call late one day, and the shop told me they had another plan. My ideas weren't the priority. Their question on that small jobs was what was important, so I had to shift. And if I left this to AI, it would have been fine, but my judgment is what made me irreplaceable here. You can't do this without a human at the helm. You can't do it well. And that's what it looks like for me. Now here's how you do it. You can do this, and you don't have to be walking a shop floor to do what I'm describing. All you have to do is be interacting with the business leaders that you serve. And if you hear anything today, I hope you hear this. I believe that interaction is the essential skill in our work. If you are not regularly talking with your business leaders, you're not much different than AI. So I want that to be a warning bell for you. Make sure you are interacting with them. And then you need to build context. Ask if you can take transcripts, save your own notes, give that to AI in a responsible, secure way, and then you need to ask what the problem might be and what you should ask next. And at the end of the day, you're gonna have to use your judgment, and that is the part that no one can replace. Alright. Back to you, Niall. Alright, friends. Our next polling question is here. But I am learning so much. I've taken some notes here so far. 95% of accounting task gonna be automated. I know we've been hearing this, but I I kind of feel like that might help free us up so that we could use our professional judgment. We can read those faces while we're having those client meetings. So with that being said, how can you use your initial conversations with a client to figure out what is keeping them up at night? Is it asking target questions about their challenges? Is it looking for patterns across multiple conversations? Using the first itch... Issue that they mentioned as their main problem and just staying there. Using your judgment to decide what needs further investigation. Inquirer minds wanna know. I know in our firm, we do ask clients what they believe their problem is. And as we're listening in those initial conversations, we oftentimes figure out that might not be the source of the problem. That just might be the pain that they're experiencing in that moment, especially since my shop is still doing compliance work. So taxes is always the main thing that they come to us for. But then you'll hear that it could be that they have a bad employee that is causing them some stress, and they never would have figured that out. But sometimes we gotta just listen and ask more questions. And, again, like Alicia said, use your judgment. Keep in mind that the AI is gonna help us listen to those conversations and pick out things, but we still have to use our HI to lean a little bit more and learn more about our clients. So, hopefully, you've answered that polling question. We we wanna make sure you're getting that CE, but we also wanna continue diving into that content. So I'm gonna, remind you now that everything that's being covered here is also in that workbook. So there are some sample questions. So as Alicia's diving in deep, you... There are questions there that you can use to in your quest with AI. There's also guidelines on transcripts and how you can navigate protecting your clients' information as well, when you're feeding it into the AI. So just keep in mind, we got your backs. Not only are we helping you teach those stories, I'm helping you tell those stories while also making sure you do them securely. And with that, I'm handing it back over to you, Alicia. So, Naya, I'm so glad you told that story because there is just this fine art. There's not a right answer to this poll question because either way, you gotta jump in and try to solve a problem for a client. And sometimes it is exactly what they're telling you, but sometimes the financial information tells you something else that they really need to know. And that's what this next section is gonna be about. So in the shop, I had my question. Right? It was about, are the small jobs taking up all of our time? And now I had to find out if the numbers agreed. So all that discovery, it gives you really good ingredients for a good story. Those conversations, they told me what people believed, but the financials, especially reconciled order ones, they're gonna tell me if what they believe was actually true. And on my own, that is hours and hours of clicking and analysis, and that used to be like a job, of days of work or maybe a data analyst job, but now I've got AI, and this is where it gets fun. So I want you to think of AI like a personal assistant. Imagine hiring the best student alive, maybe from Harvard or maybe MIT. It passed the CPA exam, the GRE, whatever exam you want. And this is an AI that it's eager to please. It wants to do amazing work for you. But here's the thing. It knows nothing about you. It knows nothing about your client. If you supervise people, I want you to think of your last hire. Even the most brilliant person, they still had to learn your clients and your expectations. And like any staff, it can have a bad day. It forgets things. It uses the wrong information, and it's gonna wander off down the path that you didn't ask for. That happens all the time. And the student from MIT, here's the other thing. It thinks a little differently than one from Harvard. So Gemini, ChatGPT, Claude, even the tiers inside each one, they all work differently. And different models, they give you different performance. And the reality is you're gonna have to fine tune it based on the task. Now I see you on this one because this is why it is so hard to navigate AI and why you might have felt like you were just getting bad results. And I'm gonna walk you through how I have navigated this, and maybe it will help you. So I wanna talk through a rookie mistake I see. It's that people use generic re... Prompts kinda like this one. Analyze these financial reports for variances. When you give a prompt like this, you're gonna get AI's common denominator answer. It doesn't know what variances matter to your business. And how many of you have watched AI analyze your reports and you thought, I already knew this? That's this. A lot of people, they blame AI. They say, that it's just not very good. But the reality is it could be your instructions because you haven't given it enough. And let's go to the other side. Accountants like to swing the other direction. We start to micromanage it. So imagine prompts like this one. Click here, then here. Download this. Format it like this. And sometimes AI needs some of those instructions, but you need to remember something. It is so eager to please you that it's gonna follow your instructions to a tee. And it's gonna limit... Or it's gonna... It will execute to the limit of what you understand. Now it knows things that you don't, and you know things that it never will. But if you over specify, you're gonna lose everything that AI knows that you haven't thought of. And this prompt is where those two meet. So there's a healthy middle, enough direction to get the job done the way you need it, but enough room for it to deliver for you. Now the middle isn't about how long your prompt is. It's about what you give it and what you leave open. Now I've got four parts that have helped me to find that middle, and this is a constantly evolving thing. But if you want your own version, I just want you to think about how you would instruct a brand new staff member who's really, really capable. If you're a good boss, you're probably gonna be a good AI prompt builder. So here's my four steps. Objective, context, standards, and checking in. Let's let's look at this. So first is the objective. What are you trying to accomplish with what you're about to ask? So my objective in the shop, it was really specific. I needed to find out, are the small jobs slowing down our sales? And you gotta know what you want before you set AI loose and give it a question to ask. And then the next is context. What does AI need that it doesn't already know? There are so many things that you can put in here for context, and I believe this is the most powerful step. You could put your transcripts. You could put your notes on your client. You could put your, business's expertise or strategic plan or your own expertise. And then you've gotta put in the right financial reports, not every report. Now for the shop, transcripts, notes, and the sales by customer detail and their profit and loss, that's where I landed. And you're gonna see in a minute that it's actually not where I started. Before it starts, you need to describe what a good result looks like. Renee Brown calls this paint done. Now there's two parts to this. There's quality and there's accuracy. And I think this is the linchpin for accountants. It's where every accountant gets nervous because if AI is having a bad day and your name is on the line, you've gotta know. And this step is how you're gonna know. So quality meant for the shop, small jobs versus big jobs. I wanted to see the small jobs by count and by dollar amount. How did we do, and how much did we earn from those small jobs versus the big? And then I had all these transcripts with what we thought was happening. So I wanted to know what's confirmed from all of those conversations. And then I wanted to know what's surprising in the numbers, but not in anything that we heard. Right? What... What's buried in that analysis that I didn't see? And then I came up with one more, alarming. What are things that we haven't even thought of that are about to be a fire alarm if we don't act today? I just loved the results from this. And then the next is accuracy. This is my account brain and many of yours. So it's saying every finding ties to the original financials, and I want formulas and check totals that I can tie out myself because I am going to check its work. And I want it to show me the evidence from its documents. So if it's making findings, I wanna know where it referenced. So it's gonna have a sentence on why each finding landed in its file. Basically, show me your work. Alright. So I wanna talk about one more rookie mistake here. You run AI and you get a terrible result, and you spend an hour fixing what I like to call and many call AI slop. Don't do this. If you can see in a few minutes that your results is off, I want you to do this. I just want you to scrap it. Just start again and plan it better. Okay. So let's talk about how you can plan it better. This is gonna help. The last step. Some people in the software call this planning, I call it checking in because I need this to go both ways. Before you have AI start, I want you to have it check-in with you. And I ask three questions. First, I ask AI, how are you planning to do this? I wanna see AI's reasoning before it even begins, which is a lot nicer than after it's done something that's really frustrating for you. Second, I ask what what do you still need? What haven't I given you? And then finally, what should I know or consider? That last question, a lot of people miss this one, but I want you to remember, AI is a graduate of a different university than you, and it's brilliant. So what if it has a better idea than you did? Wouldn't you wanna know that before you got started because it might change your direction? And that third question is how you keep from micromanaging and also how you might get results that you couldn't have even thought of. Alright. Now I wanna describe why I'm glad I did this planning stage and how AI was about to go off the rails. So when I set up the analysis for the shop, I was gonna give it the general ledger to review all the sales. I love the general ledger report because it's got everything in it. And AI came back and it said, actually, that's a really hard report for me to read. It's really long, and I might not read it well and miss things. I was like, that... That's so great. I'm glad it corrected me. And I said, well, what what do you suggest? And it said, I think what I'd like to do is look at your accounts receivables. Alright. Now if you're an accountant, you know that that's not the right report. Receivables show us who owes us, not who's buying. They may have already paid. So I had to resteer it to the sales by customer detail, and I was so proud of myself for catching it. And that little extra step saved me a ton of time and a headache. Alright. Back to you, Niall. I'm gonna let you run our next poll. Awesome. I'm learning so much here. And this is the section that I was most excited about. Right? The analysis. I love finding the healthy middle. I have definitely been guilty of micromanaging my AI. So our polling question is where do you get stuck most often when working with AI? Is it using prompts that are too generic? Is it giving overly detailed instructions, or is it spending too much time fixing poor outputs? Luckily enough, I haven't been spending too much time fixing those poor outputs. I don't know. I can always tell when my AI is having a bad day, and I'll be like, we'll we'll we'll come back to this chat. But I'm learning so much. I love the four, four step framework you gave for analysis, I definitely wrote that down. And I never even thought to find alarming findings. So that is now on my things... List of things to do here. So if you haven't already checked the poll, come back. What are you doing? We're learning so much here. Select an answer so we can get you that CPE credit and then get back to to the findings here because Alicia's teaching some good stuff. I got got my notes here and my notebook because I'm outstanding. I wanna hand it back over to you now so that... Alicia, so that you can tell us, what the discoveries looking like in your dad's shop and what that analysis were landed you? Where did it land? Now, you know, I'm I'm interested in the juicy details because I love a good a good juicy story too. I love it now. I want some juicy details because sometimes people talk about this and generically, I want to pull open the hood. I wanna show you firsthand what this looks like. Now I'm about to show you some screenshots of what I did in Claude, but I can't show you the shop's real finances. Sorry. Not gonna do it. So I recreated what I did with the made up shop. It's called Ridgeline Machine Works. The owner is Walt, and the shop manager is Dev, and the numbers are made up. The final story is a little different than my real story. It's not the real information or even the real story, but it's similar, and the steps are exactly what I did. And what you're gonna see on the screen is what AI actually gave me. And in my case, I use Claude, but everything from here until I say in real life, it's this fake machine shop, my... Not my dad's. Okay? So I made a hard call today, and I hope this is okay. I'm not gonna show you every exchange I had with AI. I had to think long and hard about this. You couldn't copy what I did anyways. There's not a formula to this. There are guidelines that worked for me, and they are going to keep evolving. And you're gonna build your own by doing this, but you're welcome to borrow mine. So instead of showing you each screen, I'm gonna slow down on what I did and why I did it. And then I'm gonna let the results speak. So this is the time I just want you to sit back. Don't worry about reading all the screens. I just want you to listen and absorb. You're gonna get all those slides and the whole workbook at the end. But you're gonna see me use AI in that four step framework, and I'm gonna do it several times from here on. Okay. First order of business, I am not letting AI do any analysis without me looking at the data first. Right? I want my own sense of things, and I needed to look at the p and l differently. Now let me just show you this. We have all seen p and l's like this. They are three, four, five pages long. And as my mom, who's not an accountant, says, it's long this way, and it's long this way. I can't easily form an opinion by looking at something like this. So let me show you my first prompt. I'm gonna read it to you. Just take this all and you're gonna see this exercise of me actually trying. So my objective, this p and l is too long to read. I want you to give me a group version that I can actually look at and trust. What's my context? I need a 12 profit and loss. It's attached. My notes explain that there was a change in the chart of accounts midyear. I want you to group by what the shop owner thinks about, jobs and materials, people and overhead. What are my standards? I want about a dozen rows, twelve months plus a total. Every number is a formula that ties back to the original document. I also wanna check row that ties the net income to the original. That's the account to me. Now checking in. I ask AI, show me your grouping before you build. Now for context, did I need the transcripts to group the p and l? Nope. Context is the right information, not all of it. So let's see it's grouping. It read all the five files, and here's how it wants to group them. Job revenue, other income, materials, consumables, labor, yada yada. I like it. Alright. So my reply, that grouping works. Give me one page. I need all the months on one screen, and then it gives me the report back. But I'm an accountant just like you, so here's how I checked the work. I click on a cell and you can see at the top, it's a formula. I love formulas tied back to the original. It is not hard coded. I love that, and I absolutely hate it when AI hands me a hard coded number. I don't want it. Then I've got my net income that it had created on the bottom. I tie it to the original document, and it matches it exactly. Now I'm trusting it. Okay. Again, I'm just gonna remind you, these are not my dad's real financials. Okay? So let's actually look. Job revenue for this fake company, it starts around 253,000 in January. And which direction is it going across that row? It's going down, down, down to about 166,000 by December. It's like about 90,000 less in in job revenue. Now let's look at cost of goods sold. Let's see expenses for this work. It comes down too, but it's nowhere near as much. So what I'm seeing right away with my own eyes is sales are down like the owner said. Now I have my own sense of it, but I can't see the makeup of the sales from that profit and loss. Right? I can just see that they went down. So I gotta dig deeper because, remember, my big question is the small sales versus the big sales. And now we're gonna get to the big Kahuna prompt, and this is the one that blew my mind when I did it. Alright. So let's look at this. It's the same four steps. I'm gonna slow down and read this one because this is the hardest task that I'm gonna give AI today. And if you can absorb this, you are going to love the results it gave and what you can do. Alright. What's my objective? I need to understand why our sales are down. I need AI to help me determine whether it's the small jobs or few big customers slowing down. So I can tell that shop manager, Dev, where he is the focus. Now what's my context? What am I gonna add to the prompt to help it do its job? I'm gonna give it the sales by customer detail export. And then I also need to define what is the small job because I haven't said that yet. So I'm gonna say it's under 2,500. I also am handing it my transcripts and my notes. I want them to read that. And then the job is to tell me what the team believes. K. Right? That's what it's gonna help it do. Now standards. What what are my standards? I want small versus big jobs. I want the count and by dollars, and then those three piles that I talked about earlier. What's confirmed, what's surprising, and what's alarming? And, again, every number tying to my original files, I want the check totals, and it's gotta show me one sentence on why each finding is in its pile. And then checking in with AI, before you start AI, how are you gonna do this? What do you still need from me? What should I know or consider that I haven't asked about? Now it came back with its plan, and I'm not gonna show you the whole thing. It was long. There was a little bit of back and forth, but it was pretty good. I read it, and I told it, go ahead because I had written really good instructions. But then right before I had to go ahead, I realized I had left something out. This whole thing, it's about this mix of customers changing. So I realized I needed to add something that I forgot in the instructions. I said, show me who bought in the first half of the year and not in the second, and then the reverse. Who is new that wasn't here before? It was gonna really help my... With my analysis. Now I thought I had set it all up perfectly in that first time, and then AI made its plan, and then I still thought of something. And that is what that check is for. Now you're not gonna get your prompts right the first time. It's not gonna be perfect. You're gonna get better at this, and it's a skill, and it's often a back and forth. Alright. So let's look at what it came back with. It came back with a lot. I'm gonna show it in a few slides. So the first is a pretty cool spreadsheet. Look at the small invoices here. It's way more of them than the big invoices in January. And then notice the direction. They have gone down all year. Now there's tabs on the bottom for all sorts of extra detail, who bought, who went quiet, who arrived. And then it also gives me pages of analysis, like, a written analysis. And it was just too much for me to show you, and it was also too much for Walt, the owner. So I asked it for the version I could read to Walt, and here is what came back. Three piles. Remember those? One finding each. Confirmed. Sales fell every quarter. Small jobs, there's 61% of the invoices, but only 18% of the dollars. So Walt, our pretend owner, he had that gut feeling about it, and he was right. Then they're surprised. It isn't the small jobs. There are fewer of them every month. The drop is the big customers. And then there's alarming. May's is down 36% since July. So we're gonna talk about her in a minute. There were other alarming things, but those were Walt's calls to make. I wanted to focus on this shop manager, Dev. He needed to make some specific calls. Remember when I asked you who bought in the first half and not the second? Here's the 10 regulars who bought in both halves. Two names stood out. There's Mays. She's the big regular. She sells sports equipment, and she makes the steel net post for pickleball courts that she installs. And that is a sport that my dad and I love to play. Twice a month like clockwork, until July were her orders, and now she's ordering about a third less. So we gotta keep an eye on May. She's the one to watch. Then there's Jeff Jalopes. He fixes up old cars, and I kinda wanna take mine down there. I have an old car. He's down too, and the transcripts told me something. He's got a Lyft rack quote, and it's been sitting on Deb, the shop manager's desk for over a month. Jeff's has an open quote. Those are the two that this data story is about. Now I want you to step back for a second and just watch what I'm doing here. Can you see how this exercise is going to help me run a business from the financial information? This is what I'm talking about that we can do. This is the kind of work that accountants are paid for. Now I've been pasting files in the cloud. I wanna show you two other tools that help you get these kinds of results with AI. Alright? So the first is the cloud connector to QuickBooks online. Now CHAD TPT has one of these two, and you'll hear more about that at the end. But this connector, it's a stable way to get your financial information into AI without having to export anything and paste it in. You connect it in the settings area that you can see here, and you choose what it's allowed to do. And it can do a lot of things. It can read a lot of the QuickBooks reports, and it can also write mostly invoices for now. But I gotta tell you, be careful anytime you are letting AI write to your accounting system. I block those until you fully understand them, and I've got some videos coming that go deeper on this, and you'll see where at the end. But for today, all I knew were two reports. Right? That sales by customer detail and that profit and loss, and both of them I could use, here. So I connect it. I give the same prompt and the same transcripts and the notes attached, and the only difference is I don't have to attach the financial reports. Claude can just go in and grab it. And when it goes to read QuickBooks, this pops up. It's asking to go get those reports. Now I... A little tip is I I like to say allow once, not always allow, because I wanna know every time it touches my QuickBooks data. And please, please, please remember those safety guidelines that I talked about earlier. This is why it's so important. But if I could just plant the seed in your mind, imagine what else you can do once your financial information can show up in your AI. Right? Okay. And it builds a report. It's not the small jobs. You got your big customers, Mays and Jeffs, call them this week. Same information, different report. It... It's... Now I want you to notice it doesn't look the same. Right? It's a different report. That's not a bug. That is the art of using AI. Okay. So the other way is inside QuickBooks itself or Intuit accountant suite itself, and that's using Intuit intelligence. And you can find it right at the top with that little icon. And look, you can see, I can attach my transcripts from my meetings right here, and I have a similar prompt. It's a little shorter for visual. But let's see what it comes up with. It gives me a pretty similar finding. It tells me that it's the big jobs, not the small ones, and it gives me a table right there by month straight from the books. So I like it because the check totals, it's all built right in. And it... It's even read the transcripts. There's Walt and there's Dev and the transcripts, it made the answer better. That's the key thing because I gave it good context, and you can do this. Okay. So I pushed I pushed Intuit Intelligence. I wanted to see could it make me a chart, and it did. I'm kind of impressed. Hold this thought. I'm not gonna get into too much detail because charts are where we're heading next. And the reality is Intuit intelligence, it's got some pretty cool capabilities, and there are ways that you could be using it today. And I've got a way for you to test it later in a safe place without being worried that it's gonna mess up anything. So stay tuned at the end. You'll hear more about it. Alright. Step back and look at what AI did. It had one job. It brought me information on a silver platter. Now a data analyst could've spent days on this pattern finding. AI put it in front of me within a few minutes, and that is a place where it genuinely saved me a lot of time. But let's talk about my job and why I was indispensable as a human in this. I took that time and redirected AI. I used it to ask more questions and decide what matters. For example, it flagged the customer that dropped off in the spring and hadn't ordered since. That could look alarming, but I was able to have context, and I went and asked a follow-up question with Walt. And I found that's a seasonal customer, so it came out of the alarming pile. My job was also to check the work. My name is on these results, so the numbers had to be right. And there's one more thing. It found what it found because I gave it the right context. It cannot bring you what you did not give it. Alright. Now I'm gonna pass it back to you. Alicia, this has been super helpful so far. And I love that you just showed us a comparison of Claude and Intuit Intelligence and how they both brought back very similar results. So this just gives me hope that all of the AI bots are, like, talking to each other even though they went to different universities. Right? I'm an HBCU girl. My AI probably went to a Ivy League. Alright. Alright. Next polling question is, how do you currently communicate findings to your clients? Are you using the reports straight from QuickBooks? Are you building custom reports? Is it through a conversation? Is it a email with bullet points? Or is it something else altogether? So right now, our firm uses the management, reports heavily within QuickBooks, but if there are anomalies or changes, we usually send a video, just because our clients, we know that they're busy. They'll watch a video before they'll go in, pull those financial statements. So just curious, what are you all doing? Making sure you answer this poll question so we can get those CPE credits. And I hope you are enjoying Alicia's storytelling journey right now. I can tell you, I I really am. I love how the AI gave you points for dev. Right? It... Walt's your customer, but but Dev's a a big part of the team, and it... You're looking like a rock star to Walt because you're like, hey. Give this to your shop manager. They're like, my accountant is the thebomb.com. I know that's the old saying for us old ebookitties. But, hey, it's proving our value here, and AI seems to be doing a wonderful job of helping you get those fees. Awesome. Just a reminder, we talked about all this stuff, and it's in that wonderful workbook. So if you haven't downloaded it, please do because it will be your resource. There are some simple prompts in there that you're gonna... You can use for this analysis exercise. There is a guide for the four state steps for setting up your AI, and then those rookie mistakes we talked about and what you need to do in order to get good, better, best results. So make sure you have that, workbook saved to your desktop so you can, you know, get to it whenever you need it immediately right away. Alright. Alicia, back over to you. I I think you have story time for us. Story time. Yeah. Thank you, Naya. And I really love that example of the video. What a brilliant way to give information. Just that... Just feels very interactive. Alright. You got all the ingredients to tell a good story, a good data story, so now it's time to tell it. And before we get going, I gotta say, I have so much I wanna say about data storytelling, and I can't. I don't have enough time. So I put a lot of it in your workbook, and it's also got some links to my firm of the future articles that you can use as resources on it. But I wanna talk big picture about three things that I think every data story does well if it's doing it well. First, it calls your reader to action. It keeps it clear, and it shows it doesn't just tell. Alright. So first, a data story has to call your reader to action. There's two things in that. It's the who and the what. Who is your reader? How do they think? How do they take in information? I want you to think about a machine shop. These are people who don't love sitting in front of a computer. So how they receive information, it has to shape how I'm going to deliberate. And then what do you want them to do? Do you want them to celebrate something, investigate something, change something? You are moving them to action. And then next, you've gotta keep it clear. We have all been handed a 15 page report with the finding buried on page nine, and it's why we don't hand our clients a big long ledger. We hand them a financial report like a profit and loss or a balance sheet. It's summarized. We can summarize even more. I want you to pick one dataset and one story, and you're gonna knock it out of the park. Now the way I do it is I like to lead with a headline, kinda like a newspaper, and the interpretation underneath. It keeps me honest about making sure I'm telling one story. And then charts, they are your superpower. People see a picture and they understand it in a way that your words cannot do. So you gotta choose the right type and you gotta keep it simple. Again, I could talk about charts all day, but today, I'm just gonna show you. So I wanna show you all three back in action, and I'm gonna be back in Claude. And here's how I set it up. We're gonna build it in two parts. Okay? So first is my headline and interpretation, and then we're gonna build our chart. It's the same four steps, and I'm asking on on this... In the whole page, headline interpretation. I want it all in one spot. Alright. So what's my objective? I need a one page data story for Walt and Dev. What do I want them to know? It isn't the small jobs. It's the two big customers that flowed. And it could get missed because new work is picking up. Now dev dot shop manager, he needs to spend his week on those two customers, not the 31 jobs on his board. Now what's the context? What am I gonna give it? I'm gonna give it the analysis that we just finished. I'm gonna give it my transcripts and my notes, and I'm gonna include one more thing, my chart rules, the things that make a good chart. I've attached that. Now, Deb, that shop manager, he runs the floor. I gotta think about how he needs information. He wants it simple. He needs big, clean, readable, honest screen that can be in a shop. And then Walt, he's the owner. He is on the go all the time, and he is reading things on his phone. So I need it short and sweet. What are my standards? I need a headline that reads like a newspaper. I don't want numbers in my headline. I want Deb's name in it and a verb and a call to action. And then that interpretation underneath, I want, like, no more than three sentences, and I want one or two numbers that carry the point. Walt has to be able to read it in ten seconds and hand it to his shop manager, Dev. And then the chart that I'm gonna put below it, I needed to show that the drop isn't from the small jobs. It's from the big ones. I also need to make sure it follows my chart rules, and I gotta make sure that the numbers tie. Right? So for check-in, you can see I've got two parts here. One is on the story and the other is gonna be on the chart, But let's focus on building this headline. So I ask it, and we can build it together. So it gave me its headline and the story right, and I... It had a pretty good idea. And so here's what it came up with on its first try. Honestly, it's pretty good. But when I read this really quickly, it's... It reads kinda more like a sentence than a headline. So I marked it up so that you can see where it fell short. So, like, in the beginning, where is the whole point about, like, sales are down? That's a pretty important thing. I think it needs to lead with that in the headline. And then Mayes and Jess Jalopes, they... I don't want customer names in a headline. It's a little bold. I don't like it. And then that interpretation, I'm already seeing it's just too long. I wanna make it three sentences, and that's where May's and Jess go. So I give it that feedback, and here's what it gave me back. And I gotta say it's pretty good. Sales are down. It's two big customers, not the small jobs. Dev calls both this week, and then you see the three sentences underneath. Now while he can read that headline in ten seconds, and Dev gets the three sentences, and we're gonna come back to this page. But let's now move on to part two, the chart. Okay. So here's the first version of the chart that it built for me. Now I'm not showing you what it, before it built this. It wanted to give me just four bars, one per quarter. And when it said that, I said, you're gonna miss all of the Nuance in July because that's where things really shifted. So this is better because of that, but it's still not perfect. There's a lot I wanna do to this. But I'm looking at it. There's things I don't love. So I use the voice to text feature in Claude, and this is me talking to it literally voice to text, and I tell it just make it better. So what is better? Alright. I like this. It's getting better. But when you stare at this, I've got... I want you to ask yourself, does this chart tell the headline? Remember, the headline is Deb needs to call two specific customers. When I look at this chart, it's about small jobs and big jobs. I needed to tell the small job story. This is important. But, also, I gotta talk about these two, Jeff's and May's, and I can't see them anywhere on this. So I wanna talk to you about a little fancy thing that you can do with charts. Sometimes you need two charts to tell one story under one headline, and sometimes you'll see them do this in the news. So I'm gonna do that too. I asked Claude, give me two charts. I asked for chart one to be this one. So sales are down, and it's the big jobs. And then you see on the right chart too, May's and Jeff's against the other regulars, first half versus the second half of the year. And here's what it came up with all on its own. Pretty cool. Now this page is getting in good shape, but here's something about AI. It gets so focused on the last thing that you asked that it forgets the thing that you started with. And do you remember the context I gave? It was my chart rules in the very beginning. So I wanted to stop and check it. Is it actually following my chart rules? But I wanted to go one step further, and I'm really excited about this one. This is a fun one you might wanna try. So my favorite data storyteller is someone named Cole Nussbaumer Naflick. She wrote an amazing book called storytelling with data, and she has great resources. I have learned so much from her books. So I said, don't only check my rules. I want you to hold this page to Cole Standard. What would she say about my chart? And I let it run. It went... Did some research, and it came back with a really big long list. I didn't wanna read all of that to you. So I just said, you know what? Mark this up. I had Claude make this for me. Mark up this chart so I can see it, and then you can see it too. Here's what it gave me. So look at this. This is a lot of what it said. There is a legend, and Cole would say, label right on the chart as much as you can. So I'm I'm gonna take that legend out. I'm trying to also raise an alarm about Mays and Jeffs. And when you look at Mays and Jeffs, there's no alarm, like, a red color anywhere. Then you've got, like, all these numbers on every row when there's only two rows that matter. There's a lot of words. And then there's, like, file names and the captions, so I had some work to do. I took most of these changes, but let's see what it came up with. Alright. One spot of red. The second half bar is for May's and Jeff's because those are the two calls, and I wanted you to see how dramatic the change is. And as I was starting this, my eye caught one more thing even after all of this work, and it's this, the point that gets buried under the dark blue above it. So if you look on the left, the point is the small jobs. Right? But that dark blue above it kinda buries it. So I asked AI to push the coloring of the big jobs into the background and put the bold color on the bottom because that's where the story is. So you could see where your eyes go. So let me show it. Do you see where your eyes can now go a little bit easier? So there it is. Two charts, one headline reading across. There's red on the two names that matter. There's bold, blue on the small jobs so your eye knows where to go. And that's a pretty tasty data story, and it's a massive improvement for where we started. And I could do even more, and you're gonna get a chance. And I I just gotta point out, I haven't even I haven't even touched this chart. I was not in Excel. This is just Claude making this for me. Let... Let's do a few more turns just for fun. So I told you why... Walt reads everything on his phone. So I asked, can you make me a mobile version portrait? And this is what it came up with. Isn't that cool? No formatting. No software. It's just right there. And then last, I asked whether we should add any of these new rules to my chart rules file, and it told me that there were three. The rules file grows every time, and now I can use it on the next client that I go. And it includes a lot of other really cool things, and you are going to get that file at the end. Those instructions are gonna be in your workbook. What I hope more than anything is that this inspires you. Because when I first saw what I could do, it blew my mind. More than anything, I want you to see what's possible and how I, as a human, stayed at the helm. Now I wanna tell you also what happened in real life. At the shop, I put a real version of a page kinda like this in front of our shop manager, and there was a real customer whose sales had slowed. And there was an open quote that had been sitting for just a little while. But because he could see the change, he saw it clearly, and he made a phone call. And by the end of the five minute conversation, we had a $50,000 purchase order that landed the next day. I want you to think about what that is. That order was going to go somewhere. Because of this work, it came to us instead of another shop. We got more sales. We got in front of customer opportunities before they went to someone else, and the business got better because an accountant was in the room. That is so much more than just getting the books in order. That is irreplaceable work, and I want all of us doing that kind of work, and I know you can. Our clients can fix things when they can see them, and data storytelling helps them see. Alright. Back to you, Niall. Alright. Our next polling question so you can get those CPE results. Which storytelling skill would you most like to strengthen? Is it calling the reader to action? Is it keeping it clear and simple? Is it showing, not telling? And I think Alicia did a wonderful job here of giving us a way to surface insights to clients in a way that's digestible to them. I never would've would've even thought about a headline. It's bold. It reminds me of something like, hey, newscaster y. Right? Where somebody can just look at it and say, oh, this requires me to do something. Especially adding a name and a verb, That was super smart. And then the idea... My clients are using their mobile phones heavily, so making your visuals, meet the place where your customers are at. Oh my gosh. Alicia, that was just so fantastic. So thank you for sharing that with us. But which skill are you all looking to strengthen? I can tell you I'm working on keeping it clear and simple. That's my favorite way to communicate. But I know sometimes us accountants get a little technical and long winded, and I'm guilty of that. So b is my option. What's yours? And remember, that workbook is going to be the key to unlocking how you tell better stories with AI. So be sure to download it if you haven't already. It's just got filled with so many good nuggets that are gonna help you. So I know Alicia has some final takeaways for us, so I'm gonna turn it over to you, Alicia, to end us on this journey of telling stories with AI. Thank you, Nao. So I hope you capture from this. This was not about AI making me faster. It was my detective. It was my researcher, my analyzer, my visualizer. But I was in the driver's seat, and sometimes it slowed me down and helped me go deeper and end up with a richer data story. And when you hear accounts talk about AI, we talk a lot about faster. I want you to go richer and deeper. And we're all learning this together. These tools have not been around that long, and they are continuing to change. So nobody's quite an expert, which means you can start today wherever you are. All you have to do is be curious. Alright. So this is your time now. I wanna talk about what's next for you. So I have a pet peeve about webinars. Often, I hear a lot of great ideas, but you don't know how you can use them and put them into practice. And we all have real businesses that we're serving, and we can't always practice on them, or maybe we don't have enough businesses to practice on. I wanted to give you a safe place. First, I want you to complete... We have a case study in the workbook, and I'm pretty excited about this one because who knows? Now do you know who's Craig the landscaper? He is the friend that let us learn QuickBooks on his company. It's a free sample company that anybody can open to learn QuickBooks. Well, guess what? Craig has a big decision. He wants to quit his day job and be a full time landscaper, and your job is to help find his story in the numbers and tell him. And you can use any AI you want. You can also do his exercise using Intuit intelligence in the platform. Now there may be a surprise or two in there. You're gonna find him. Now second, I've got a spot on my website with all sorts of free resources. I'm gonna be adding practice sets for different types of accounting work. I'm also gonna show you how I did this exercise in Gemini and ChatGPT and Copilot, and I'll have some more probably on Ridgeline Machine Works. And then I've... You can follow the shop story on my LinkedIn. I'm gonna have lots of fun juicy stuff that's... I'm talking about there. And then last, I am starting a Facebook group. Now I have to admit, I love Facebook. I don't know if that ages me, but I love using it for all sorts of things. But my favorite part is it has groups, and accounts need to talk to each other. I love Facebook groups. So this one's called the human accountant figuring out AI together because that's kinda what we're doing. So you can bring your questions, you can ask questions about this, share what you're trying, and then we'll figure it out together. Alright. Lots of great opportunities and one really exciting one that's coming up, and I'm gonna pass it to Artie to share more about it. Go ahead, Artie. Amazing. So we are super excited. We are going to have another meeting where we come together on December 3. It's on Thursday. You'll see all the information on your screen. There's a registration link in the slide deck, and there's also a questionnaire. So the questionnaire is going to ask you a few different things about what did you do, what did you try, what worked, and it'll even give you a little spot where you can upload some of your work to show your work. And then, also, if you have a question for Alicia about something that you learned, something that you did, something that you heard today, she's gonna go through everything and we'll come back with a bunch of questions and examples and even kind of call out, you know, who shared what with us. And, really, this is just a way to kind of bring everybody together and keep us in this community similar to how Alicia's gonna have the Facebook group, but allow us to come back together so that it's not just a webinar where you're taking one thing and hearing one thing, but we're allowing you to come back and share what you did with your peers. So we're super excited. And, again, just another reminder, all of these details are in the amazing workbook that Alicia created. And so with that, we are almost done and we've got one more polling question. And we just wanna know how did we do? So how likely are you to recommend this webinar to a friend or a colleague? We're super excited that this is going to be on the ProVisor Academy, so you can access it through your accountant suite or QBOA, and share it with friends. Go ahead and forward it if you'd like, but we're really excited to hear kind of your guys' feedback. And then with that, just a huge, huge thank you to Alicia and Nio. Alicia, you have put so much effort and work into this. I learned so much and I'm not even an accountant, but now I'm excited to take this into my world and use your examples and your four steps in delivering things that I need to do too. So thank you so much. And, Nio, thank you so much for coming and joining us. Love getting your perspective on everything as well. And just a big thank you to everyone that attended today. We're so excited to share more. We wanna... We want to hear your feedback, and, hopefully, we can continue to do more of these sessions if you find them valuable. Thanks so much, everybody. Thanks for having us. Thank you.