← School Business Insider
School Business Insider
Beyond the Buzzwords: Real AI Solutions for School Business Officials
01:01:35
Loading summary
Transcript109 lines
- [00:01]
John Brucato
You're listening to ASBO International's School Business Insider. I'm your host, John Brucato. Each week on School Business Insider, I sit down with school business officials and industry experts from around the world to share their stories and explore the topics that matter most to you. Find out what it means to be a school business official and get your insider pass on all things school business. Hello everyone, and welcome back to School Business Insider. Artificial intelligence is reshaping industries across the board and school business operations are no exception. From budget forecasting to meeting summaries and data driven decision making, AI has the potential to enhance efficiency and productivity. But with new technology comes questions, challenges and ethical considerations. Joining me today is Aziz Agaev, CEO of Flowist, an expert in AI driven solutions. Today, Aziz will share insights on how AI is already transforming school business operations, the five phases of AI adoption, and how SBOs can move from frustration to productivity, how AI can assist in budgeting compliance and stakeholder communication and ethical concerns, job security and responsible AI usage in school districts. If you're wondering how to integrate AI into your workflow without feeling overwhelmed, today's episode will provide practical strategies and insights to help SBOs use AI effectively and efficiently. Aziz, welcome to SBI. I'm so happy to have you today.
- [01:37]
Aziz Agaev
Thank you. Thank you. Glad to be here.
- [01:40]
John Brucato
Absolutely. So why don't you tell me and our audience a little bit about you, what flowist does and really give us an overview of how artificial intelligence is transforming school business.
- [01:50]
Aziz Agaev
Sure. A little bit about me. I was a school business official for 14 years for schools. And then after I left, I continued working with school business officials. So for the past 18 years, I've been involved with school business. During my CFO years, I always did automation. I did workflow automation with Google Platform and wrote a few little scripts to automate our workflows for our schools. And that kind kind of was my hobby. I decided to change my hobby to my job and now I do full automation and AI trainings for. For school districts. Now with Flow List, I go from district to district. I train their staff on how to use AI. Today, basically I give them use cases instead of just talking about big words. Right.
- [02:41]
John Brucato
That's great. Well, that's. That's wonderful to hear. I didn't realize that you were a former cfo. School business officious. So you're really speaking from experience. And it's so nice to hear that you're giving some practical solutions because I think a lot of the trainings and webinars and PD that we have seen from our seats is very much theoretical in nature and not so much how can you apply this to your day to day job? So that's great.
- [03:05]
Aziz Agaev
Exactly. Yeah, that was my approach. Right. I actually started about six years ago with Excel trainings. And all the Excel trainings we would attend would talk about apples and oranges, or we'll talk about sales data and then we learn the formulas. But then when we come back, how do we apply those to our job? That was my approach. Like what can I show you today that you can go back and use today instead of relearning it with your data?
- [03:35]
John Brucato
That's great.
- [03:36]
Aziz Agaev
And this is exactly same thing I'm doing with AI, Basically giving use cases of what problems did I have as a school business official so that you can solve those problems with the AI.
- [03:48]
John Brucato
So I have to say it's pretty atypical for someone to be in a school business official seat for so long and then kind of jump careers because most of us are lifelong school business officials. So what really motivated you to kind of jump ship to doing your own thing with AI?
- [04:05]
Aziz Agaev
Well, I would like to say thank you to my wife for that first and foremost. Yeah, I was, I was a CFO for, for in Boston area and she got tired of snow and she said, let's move. And I said, okay. And then she said, let's just quit all your jobs and move. And I said, yes, okay. And then basically that was the decision that we moved from Boston to Dallas and I decided to quit the CFO job again. My hobby was that and my passion was automation. And I enjoy doing automation a lot more than, you know, looking at number and writing checks and approving purchase orders. So I decided, you know, let me try this. And as you can see, I like talking too. So I decided to just go around and train people on technology. Because what I realized in education, we always lack with technology. Right. We always follow behind. We always come like 10, 20 years behind. But this AI technology is completely different. It's not like a new technology that is out there that we can now learn. It is more that you can ask AI how to use AI. It's completely different technology. We don't have to wait.
- [05:28]
John Brucato
That's great. Well, as a native Northeasterner, I am jealous that you got out of the snow because we had a lot of snow up here this year. So every year I'm like, why do I do this to myself? But you actually took the plunge, so congratulations.
- [05:40]
Aziz Agaev
Exactly. Yeah. Yeah.
- [05:42]
John Brucato
So tell me, what are some of the most Exciting AI applications that you see being used by school districts right now. I mean, it's kind of an ever evolving technology and there, I'm sure are many use cases. But what are you most excited about right now?
- [05:56]
Aziz Agaev
Now what I try to focus on is only a few things because every single day there are tons of AI applications introduced. There are a lot of AI machines the the LLMs and there are a lot of tools for, you know, used for photo and video and you know this, that it's very polluted market and it's really hard to go through. So in my trainings, what I try to focus is how do you use what you have? One would be ChatGPT, other one would be Gemini, Google, because most school districts are Google districts so they have access to Gemini. With ChatGPT and Gemini we can accomplish, I would say 80% of what we are trying to do. But then I also talk about NotebookLM, which is kind of podcast creator and one cool use for that would be schools can take NotebookLM and use it for their documents. And I joke about it a little bit that all the documents that we send to our employees, no one reads them. So create a podcast and send it to them and then they can listen instead of reading. But with ChatGPT and Gemini, we can continue talking about use cases during our conversation. But the main job can be a lot more efficient with only using these two. Even if you don't use anything else.
- [07:35]
John Brucato
Do you find that a lot of your training and a lot of use cases are with LLMs? Are there other AI technologies outside of language models that you see applicable to school business? I mean, I think, you know, we're a Microsoft district and I noticed that Excel has some AI plugins now that you could use. Are you seeing other technologies like that kind of being integrated? I mean Copilot is one, but outside of Microsoft, are you seeing other products that aren't large language models as use cases?
- [08:06]
Aziz Agaev
There are a lot of, you know, different large language models, which is Claude is one of them. Grok was recently introduced. Deepseek is recently introduced. It's just outside of the U.S. that's why people are hesitant. All of those models would do would be very similar to each other from our perspective. Yes. One of them would be a better code writer for software engineers. Other one can be a more writing skill better. One thing I want to mention is perplexity. Perplexity is a good one for school districts, districts because it comes with sources. Now I'll give you two use cases if you ask Chatgpt can you research the AI usage in education industry? It would give you answers, it'll give you, you know, 50% of schools use AI and all that. But then you ask, okay, where did you get this information? The information it got was from general knowledge that it was trained on. But then you can say, now I want you to bring me resources, bring me sour that you got that information and it would go to the Internet, search the website, you know, all different websites and then bring you the information, this time with sources. Basically perplexity is that it's a Google search with AI.
- [09:37]
John Brucato
Okay.
- [09:38]
Aziz Agaev
Which can be very, very interesting for school districts.
- [09:44]
John Brucato
So you know, a lot of the way I see it being used now, it's you go to OpenAI's website, you go to Gemini website and you use this kind of in addition to do you see these products being integrated into everyday use? So it's a little bit more seamless and not just an extra thing that people are having to go to. Tell me, tell me what you see kind of on the horizon for that at this point.
- [10:07]
Aziz Agaev
No, because that needs a lot of integration, that needs a little bit of knowledge of creating agents. Basically that's what you're referring to would be an agent that wakes up in the morning and does things for you and then, you know, reports to you kind of thing that is becoming easier and easier on a daily basis. Because there are a lot of platforms like Zapier, make.com, you know, these, these can automate tasks. But at this point the Gemini is, is a little more leaned towards that because Gemini works with the workspace. One of the use cases would be if you ask Gemini, read my emails from today and summarize everything and create tasks out of it and then you can say Google Tasks. Create tasks in my Google Tasks. That would be a one stop shop because Gemini would access your email, would access your Google tasks, or would access your Google Calendar and give you a seamless response.
- [11:10]
John Brucato
Great. So how do you see AI improving school business official efficiency and making data driven decisions? I mean it's something that is really inherent in our work, but I see AI as just an enhancement to that. So talk me through how you maybe see those efficiencies increasing.
- [11:26]
Aziz Agaev
I can give you a few examples. One of them would be creating new policies. Creating new policies would involve a lot of parties, a lot of people would write parts of the policy and then it would be compiled to one document. Instead you can just chat with AI, give your bullet points, give your context of the district and create the policy you want or states sometimes share the sample policies. You just take the sample policy, provide the context of your school district. And when I say context, the context is basically an information available publicly about your school district. It can be in Department of Education's website, it can be your report cards from again, doe, and it can be your own website. Whatever information is on your website, you compile all of that, either providing the link to ChatGPT or Gemini so that it learns. And then at the end you say, okay, whatever you learned regarding my company, can you summarize that? And when it summarizes it, you just take it, copy it, and create a PDF file. Now, next time you interact or anyone else interacts in the district, they can upload that information. And now you have the context of the company, basically taking the sample policy, making it your own. I'll give you another example for this is, this is for superintendents, right? Every superintendent has the evaluation criteria. They have KPIs, they have metrics that they evaluate their staff. These metrics can be improved with AI. And then the templates could be created with AI. And then superintendents, once they have their meetings, they can upload their notes regarding an employee. And it would evaluate that with the KPIs that it created. And then what superintendents can do, they can upload every single note and compare their staff to their staff's performance. You know, how each, each building is doing, how each principal is doing. From, from the business office perspective, there are lots of, lots of uses, but we can talk about a few. For example, if you have expense data for, for past year and this year, that can be a 20,000, 14,000 line expense data. You can upload both files and you can say, evaluate these files and find me patterns. If there are patterns from one year to another year, do I have increased vendor expense? Let's say it would and I tested it. It's not. Again, I want to emphasize that it's not perfect. It will be perfect. It's not perfect at this point. However, if we start doing this on a daily basis, we will be ready when it's perfect. We cannot wait until then. It would give me an information saying that Amazon, you did only 300 purchases last year, you've already done 800 purchases. So that's a pattern that you need to watch. Your cost for supplies increased from 100,000 to 500,000. There's a huge jump. There's that $400,000 jump. What is that? Right? And it doesn't, you don't actually have to say that what I would do, I would upload these two data sets and I would explain what it is and then I would say suggest what kind of analysis I should do. And then it suggests, you know, it says the pattern analysis. Then you can say year to year comparison, all of. Then I decide on which one I want. I say, let's start with this. You do it and then it starts analyzing it and it gives me the information. Basically it's a back and forth process. It doesn't happen overnight. That's why I have those six phases of experiencing AI. I don't know if you've had a chance to look at that, but it's basically the first time is surprised, then you become curious, then you're confused how it does it. But then you get frustrated because you cannot get what you want. You know, you want to ask question, cannot get the answer. That frustration part is the lengthiest part of this experience. But then you know, you have joy because you start enjoying what you do. And then productivity comes last. So it doesn't come overnight, but with a few steps you can do it right.
- [16:07]
John Brucato
So there still is a need for how to use AI and human intervention. You can't just give it a few data sets and expect a perfect response. Although you indicated that maybe sometime in the future we will be at that point. But for now you still need to make sure that you're looking over the data and having some kind of analysis of your own to ensure that the results that it's spitting back out to you are accurate and you're not kind of being misled. If AI potentially is misinterpreting what you're looking for, is that correct?
- [16:38]
Aziz Agaev
Exactly, that's exactly correct. And I think the most important thing we need to focus is to use it for our personal life. Because when we start to use it for our personal life, then we will be ready for our professional life. Even if we don't have access to AI in our school districts, we can use it for like if we. I'll give you an example. If we have Gmail account, if we have Google account, the AI is Gemini, is available for everyone. Extensions are available for everyone. You're traveling, I come to Virginia, let's say and I am looking for a restaurant that I can eat in. I can just say at Google Maps, it's a trigger that opens Google Maps. You say at Google Maps find me a restaurant that serves good steak with five star reviews with this in the review. And that's like a day to day AI usage, right? Or you can say at Google flights, find me the cheapest direct flight from Dallas to London. And it would just search the Internet and it would give you the best option because it does analysis itself.
- [17:56]
John Brucato
So it really is just kind of having the world's most intelligent personal assistant with the most resources, right?
- [18:01]
Aziz Agaev
Exactly, yes. That's how I look at it.
- [18:04]
John Brucato
So I've had guests on the podcast in prior episodes, and one of the main themes that they talked about is the importance of how you prompt artificial intelligence.
- [18:14]
Aziz Agaev
Exactly.
- [18:15]
John Brucato
Is that still relevant? Is that still something that needs to be learned by humans on how to appropriately prompt? And do you see with the improvement of AI, maybe the technical aspects of prompting are less important? As AI becomes more accustomed to how humans interact with it, I think it.
- [18:34]
Aziz Agaev
Will continue being important. It'll be just less complicated or, you know, it'll be probably a little bit straightforward. But structuring your prompt is, is, is huge. And that frustration part that I spoke about, that frustration can be a lot less, a lot shorter if you know how to prompt. But when we say people need to learn how to prompt, that is different than anything else, because you can learn how to prompt with AI. You can ask a question saying, that is my prompt good? What would you change in my prompt? Or you can say, what is a good prompt? What is a bad prompt? It'll give you a bullet points. You read those bullet points and make sure you apply those in your prompt. But a rule of thumb is it's very wide in the Internet is to start with act as or assume role. So you say assume the role of school business official and then you give a context of your school district. That is very, very important. Right. We assume that the AI knows everything. No, it doesn't have prior knowledge. It doesn't know who we are. So we give the context of who we are and what the company is. Then we provide what we want to get out of it. We can provide the design format. How do you want the design to be? For example, if you want an email, then say email. If you want a LinkedIn post, say LinkedIn post. If you want a newsletter, say newsletter. And it will reformat that way at the end. The most important part is ask me any clarifying questions before you give me a response. This basically eliminates all the assumptions. You know how they talk about hallucinations, they talk about assumptions, right. AI assumes every single part. If you don't tell it to ask. All the assumptions now are a question. And if you answer the questions correctly, it would give you a Response that you can take and use so you.
- [20:47]
John Brucato
Get a more refined and arguably accurate answer, rather than letting AI just kind of calculate the probability as in what the assumed exactly. Aspects would be.
- [20:57]
Aziz Agaev
Exactly. Yeah. Dr. Godwai Saba Godwi. She has spark. You know, she talks about spark. That's another kind of way of interacting with AI. There are a lot of different kinds of interactions. The most important thing is giving context who you are and then letting AI asking questions and assigning a role to the AI, because then it will look from that perspective.
- [21:32]
John Brucato
So really it's the old adage, you know, garbage in, garbage out, but the more you put into it, the better product that you'll get back in terms of a response.
- [21:40]
Aziz Agaev
Exactly. Yes, that's right.
- [21:43]
John Brucato
So I think sometimes we. We feel that AI, although it's the kind of the next wave in technology and is at the forefront of our minds, that we may think it's everything to everyone, but I don't know that that's necessarily true. Can you talk to me about some misconceptions with artificial intelligence and maybe what it can't do for people?
- [22:04]
Aziz Agaev
Now, one thing I would say is that there's this huge issue everyone is talking that AI will take our jobs. AI will take our jobs. And I give this analogy usually in my trainings. My dad was brilliant engineer. And when he. When during his, I think, 50s, the autocad, which is an engineering software, you know, to draw the blueprints, the engineering software was introduced and he decided not to learn it because he was really good and he was in huge demand. Right. In his time. What happened after five years? He lost his job and now he's looking for a job and he couldn't find a job because after five years, everyone required AutoCAD as part of the engineering position.
- [22:53]
John Brucato
Sure.
- [22:54]
Aziz Agaev
Used to be that they would hire two people, one youngster that knows AutoCAD and one senior engineer that would talk to that Youngster for the AutoCAD. But after five years, it is now a requirement. Similar thing comes to AI. Now, it's not that AI will take our jobs as school business officials or school personnel. It's that people who know AI and who utilize AI on a daily basis will replace our jobs. That's how I look at it. Because AI cannot do everything at this point. And maybe we have five to seven years for that to happen. But until then, let's enjoy the ride and let's learn and keep up and make AI our assistance. That's. That's how I look at it.
- [23:41]
John Brucato
Right. So I Mean, you could just even go back even further and make the argument where if you were a school business official decades ago and used ledger paper and then everybody migrated to an ERP system or an electronic means of accounting, there's. That's where the industry's going, Right. So if you had the skills that exactly allowed you to use those systems, you'd be more marketable than somebody who was, you know, scribbling in a piece of paper, so.
- [24:05]
Aziz Agaev
Exactly. And that, that's another analogy that I talk about. You know, if you don't know computer today, would you get hired? But would you get hired 50 years ago? Yes, you would. The only difference between AI and the other technology is that AI is fast paced. It improves in a huge speed. It's not like computers evolving and people getting used to computers. If you can start today, interact with it and be a lot more efficient than you were yesterday.
- [24:39]
John Brucato
Absolutely. So you were one of the inaugural articles on School Business now, the new site that ASBO International has released for all articles contributing to School business. You talked about the phases on this podcast, but you really kind of dove into those into your article, and I believe I have them. Correct. You said surprise, curiosity, frustration, joy, and productivity. Can you dive a little bit deeper into those and what those actually mean? And you know, I, as you were kind of touching on it earlier, I was thinking back to me first using AI and getting frustrated. Well, this isn't giving me the output, but talk me through those other phases and really what that means. If you're just kind of dipping your toe in the water of artificial intelligence.
- [25:21]
Aziz Agaev
What I see from my trainings, and I do for a lot of school business officials and different school personnel that even who thinks they use AI on a daily basis, they are just using the surface of it. They're just touching AI. They're not using AI. When I say curiosity, we have probably a huge percentage of our school staff that are still curious. They want to learn about it, but they're scared because they hear a lot of big words. They hear these terabytes, petabytes. They hear this GPUs and all that lingo. Right. That we are afraid to touch. That's why I don't talk about any of those during the training, because that scares people. Right. We are at that initial curiosity stage at this point. For us to get into the frustration, we need to start using it. The confusion phase that I talk about is basically when you start using it and you're surprised, you don't comprehend how it does, what it does however, it used to be doing that for a very long time. I'll give you an example. Like Google Maps, everyone uses Google Maps on a daily basis and they don't think that they are. It is AI. It is AI. Google Search is AI. You know, when. When you start typing, it gives you suggestions. That is AI. The facial recognition of our phone is AI. The people don't realize that they use AI on a daily basis. That's why they think that AI is hard.
- [27:15]
John Brucato
Right.
- [27:16]
Aziz Agaev
Basically going through those phases. And then joy part, I want to touch upon a few things on the joy part. Joy part is when you start enjoying the AI usage. And I do enjoy because it creates. It's a huge time saver for me. But there are a lot of tools that you don't have to just increase your efficiency, just make it fun. There's this suno.com for example. It's a music AI that you can create music. Now, looking from my personal view, I can just create fun music regarding my friends and send it to them and make their day, which I do a lot. And I try to mess with people sometimes too. But on a professional level too. If you write a music for your teammate, what will that do to your teammates with their name in it, with what they did? I'll give you a use case for that. One of my clients in Chicago, she had an issue with plumbing. The pipe burst and it was during the weekend. Her maintenance team came and fixed without any issues. And then next day she wrote a song for her team, mentioning every single one in the music. And you should see their reaction. They immediately responded with thank you. Now this is a touch that we can add to our life that is easier. Now that's what my approach is. And that's the joy part. Right. Enjoy doing this.
- [29:07]
John Brucato
It's like writing somebody a thank you card. But now it's so much more personalized and unique, right?
- [29:11]
Aziz Agaev
Exactly. Yes. Now it's personalized. Even thank you cards. Right? When you train one of your AIs to just write thank you notes and you then say, okay, I want this thank you note to be for Aziz and I want this to be a friendly note. You just say Aziz, comma, friendly, comma, I didn't like what he does. And then it would generate a thank you note. You know, if you want a thank you note for John, you say John, great podcast. And so. And so. Right. And then it generates the thank you note for that. So kind of making it fun and enjoying what you do is part of the job and part of the journey.
- [30:03]
John Brucato
And so then productivity obviously seems to be that last step where you enjoy using the product, but now you're using it in such a manner that you yourself are being more productive. Tell me more about that productivity stage. I mean, is that really kind of where the light bulb goes off for school business officials?
- [30:20]
Aziz Agaev
It is. And the reason I did productivity last because when you enjoy what you do, productivity comes seamlessly. You don't have to force just comes. But productivity, it can increase the value of our time drastically. I'll give you one example. There was this school district and we're still working on that project. But the school district has about 152 copiers, their accounts payable every time they receive the invoice from the copier company. They need to hand key every single serial number and how many black copies are billed and how many color copies are built. The reason for that is the build doesn't have which accounts we're assigning it to. And we are still working on it. You know, we're still training AI, but we fed the PDF. If the PDF was shorter, it would spit out Excel spreadsheet with every single copier, serial number, the black copy, the color copy, and calculated cost. Now, if that saves them two hours a month without mistakes, that creates an efficiency. There's another. We're working with Chicago Public Schools right now on some automation projects and looking at the timesheets they enter timesheets once they receive the PDF, PDF timesheets they enter. So when, when you calculate, it's about maybe 10, 15 hours every two weeks, which is, which is huge. And we can create AI assisted processes to do that. Basically those small touches would increase efficiency or from, from the business manager's perspective. Business, business officials write bids and sometimes we don't know what specifications to use, what, what to write in the bid. Just have AI's assistant to create the specifications, right? Write the whole bid, create the structure, because it can do it and it will do it a lot faster than we do it.
- [32:39]
John Brucato
Right. So what do you think are the biggest barriers to school business officials getting into the AI space? What is preventing them from really taking that next step?
- [32:49]
Aziz Agaev
One thing, and the first thing would be mindset. We need to change our mindset and we need to embrace this new technology instead of being scared of it, embracing it, and start using it. The second thing is privacy. Now, there are a lot of ways you can go around privacy. For example, if you want to analyze the employee data, you don't analyze employee data by uploading your staff's names and their addresses. You redact everything you know, you replace all the names with random ID numbers and then you assign those ID numbers to your staff. That way you can upload any information you want. Because now there is nothing that AI can use to identify people. Again, same thing with student data. Don't use the student IDs for your school district. Just create random IDs for each student. Now you can upload the student data and analyze there. So another thing is the privacy. But there's this good analogy that Mary Ellen Norman from Boston, you know, she's a colleague, a senior business official and colleague, and she talks about how do we use Google search. Right. It's very similar to Google Search. If we're using Google search without hesitance. Right. We are not hesitant to use it, then what holds us to use AI?
- [34:23]
John Brucato
Yeah, Google Search is almost muscle memory now, right?
- [34:26]
Aziz Agaev
Exactly. Yeah. And we don't say, oh, if I searched, what would happen? Or if I search, this would become a public record. Yes, it may. But you're not searching wrong things. Right. You're creating efficiencies. So that's how. I mean, I think it starts with the mindset. We need to embrace it and we need to. And I do this in my trainings. I say, train your subconscious mind that this is easy every day. In the morning you wake up, you say, AI is easy. I will use it today because it's that easy.
- [35:03]
John Brucato
Right, Right. Are there any other small steps that districts can take to introduce AI without overwhelming their staff? I mean, we all have enough going on. I think your anecdote about waking up and saying that AI is easy to use to really get your subconscious into the flow of doing things. But what else can school business officials or districts in general do to kind of maybe just take some small steps to introduce AI into their culture?
- [35:28]
Aziz Agaev
I think the trainings are very important. The reason I say trainings, when I go to a training for a school district, there are a lot of sparks of ideas. Right. We talk about use cases. The use cases are the most important part, the crucial part of this training. Because if we don't know what we can use it for, then it's impossible to use AI because we can't think of use cases. Now, I would say there should be an AI ambassador in every school district. That AI ambassador can do research and kind of make it as. As consumable as possible for the other staff members and would do weekly monthly trainings on, oh, this is what I found. You know, share best practices. This is what you can use for. For. And the problem is everyone in the industry will be talking about teachers and students, central office staff. We will always be left behind because we kind of are hidden. I call them hidden heroes. Right. No one knows they exist, but if they didn't exist, the business wouldn't run. I think that that's the most important part of finding out use cases and introducing those use cases to staff instead of saying AI is good and it can create efficiency. How it can create efficiency and why do we need that? Efficiency is the most important part.
- [37:15]
John Brucato
Great. So Aziz, can I share with you two use cases of mine that I've had recently?
- [37:20]
Aziz Agaev
Sure.
- [37:22]
John Brucato
One is very basic and one saved me a lot of time. So the basic one was actually just today, 30 minutes before you and I hopped on to record, I was finishing up my long range financial plan that has very complex formatting. It's 30 pages long. A lot's going on in that document. So historically what I would do is I would proofread it myself. I'd have my, my administrative assistant take a look at it. I converted it to a PDF and I uploaded it. Prompted that. It's a master proofreader and editorial wizard. Can you proofread this for spelling, grammatical and then just really formatting? And it gave me a section by section comprehensive list of just anything from a missing comma to doubling up a word by accident to really restructuring sentences and simplifying what the complex concepts I'm trying to get across are trying to say. So that saved me a lot of time just by reading and reading and reading. And you know, I'm sure you authoring articles, when you read your own stuff, you miss that because you, you're reading it in your own voice. Right. So that was huge. That, that was a big time saver. And it took all of, I don't know, four seconds to spit out and start kind of giving me those examples. So that was amazing. The more complex solution that I found this was back in the almost a year ago, I was in negotiations with my teachers union and two things we always negotiate time and money. And money was on the table in our teacher's contract. We have a really extensive list of coaching salaries, co curricular stipends and things of that nature. And the teachers were claiming that when comparing my districts to neighboring districts, we were much, much lower in terms of what our coaching stipends were co curricular. All those extra bells and whistles that teachers get for extra work. And you know, I think anecdotally I didn't disagree. But I didn't know that everything we were way out of line with everybody else. So I said, all right, give me three other districts that you say are paying well in advance of what or exceeding what we're paying. So they did. And you know, all these contracts are publicly accessible. So I downloaded all the PDFs and then I downloaded my district's contract. I put them all into chat GPT and did a series of really specific prompts. And basically what I said was, look at these charts, find comparable titles and compare the salary and give me an output of where my district is in relation to those three on average, whether we're below it kind of right in the middle, or we pay more. And it did just that. So it gave me a comprehensive table that I could then convert into Excel to show me that these are the stipends that I'm underperforming, where I'm kind of just in line with everybody else and maybe where we pay a little bit more. And had I done that manually, it would have taken me a couple of days probably to really sift through four individual contracts trying to match up coaching titles and everything and making sure that my data was accurate. Now, to our earlier point, I had to go through and kind of spot check and make sure that everything was legitimate. But I have to say, Aziz, that was such a time saver and an incredible use case for me of using artificial intelligence effectively that would have otherwise cost me a lot of time and effort to try and get to that ultimate conclusion. So that's very beneficial, I would have to say.
- [41:01]
Aziz Agaev
Exactly. And those are the things that you say. Okay, now I'm hesitant of paying $20 per month, but how much would it cost for me to spend three days on a research? Three times. Three days, times your hourly rate. And now tell me the calcul right now. It became a lot easier. Now with deep research, what you could do now, if you wanted to compare yourself to other districts, you can just do a deep research with ChatGPT and it would go find those contracts, hopefully the right ones. Sometimes it finds the wrong ones, but it finds the contracts. It then creates those summaries and it creates the comparisons and then spits out, even if you don't have to look for them right now from the time you used it because the deep research was not there to now, maybe two months now, Deep research is here and it can do things that human would spend hours and days on the research and it does pretty good job. I'm using it on a daily Basis.
- [42:10]
John Brucato
It's funny you mentioned that because again, when I was doing that proofreading function earlier this morning, I noticed the deep research button and I wasn't familiar with it. So this is great timing. I'm glad I brought you on. So now I know that's another tool in my toolbox.
- [42:24]
Aziz Agaev
Right, exactly. With deep research, it basically goes and does the research for 20 minutes, 30 minutes. It reads a lot of websites that you didn't think that you need to go. That's what happens to me. Oh, I say, oh, this website exists. I didn't know that. Right, right. It just goes, does the research. Then it, it basically writes you a paper on how to do what to do. It cites every single thing it took from the Internet. You can just click and go and read yourself, and it creates those comparisons. I'll give you another example from one of the districts they did. This is my school district and this is my background. Find me other school districts that are in exemplary list of the state. I want to go visit them. But make sure those districts are the same background as me. Right. That by itself would probably take two, three hours just to find the school's name from the list. Right, right. But it, it probably takes you two days to just, you know, research their background, their demography, why they're successful and all that. It just does it for you. Another use case for the business officials, I would give you one. We all create budget books to share with our stakeholders. The budget books, we have to have tables and graphs. Then we have to have narratives. Narratives. If we write them, sometimes we have bias because we understand what's going on. Stakeholders, on the other hand, they are parents, they are teachers. They don't understand what's going on with the numbers. What you can do when you create your budget book, you just screenshot your tables. Let's say you screenshot your revenue table. Again, I want to emphasize this. First, you feed the context of your school district because it needs to know who you are. Then you just screenshot your revenue. You put it in there, and then you say, write me a narrative. It writes you a three to four paragraph narrative. You can limit it, you can shorten it, you can shape it any way you want. It writes you a narrative. Then you just take that narrative, put it under your table. Then you have a graph that you generated. You take a screenshot of that graph, paste it in ChatGPT or Gemini. They both do it and say, write me a narrative. And it would just take that image. It can be a pie chart, a Bar chart. And it breaks it down to. From the stakeholder's perspective. Now, I'll give you one more thing. If you say, make this ESL friendly, if your district is heavily ESL district and you know that parents don't, you know, the English is not their first language, now you can say, make it ESL friendly and it will lighten the language to make sure that ESL people understand. Basically, there are tons of things that we can use. I'll give you one more use case. The translations. The communication is the key, right? The communication is our pain. We cannot communicate with our parents in different languages. But now we can. What we can do, we can translate one of our newsletters to 50 different languages. We can give this to 50 different speakers just for them to proofread. And now we save tons of, tons of money because proofreading, you can pay probably $10 gift card for proofreading, but for interpretation, you will probably spend like 200 to $500 for translating one document.
- [46:23]
John Brucato
Yeah, I love that idea. I love that idea of screenshotting a chart or a graph and creating a narrative. And I know exactly what I'm going to be doing once we hang up with these. There's a lot. There's a lot of complex data. And I think, you know, it didn't click until you were kind of walking through. We do have inherent bias because as school business officials, we're in the numbers. We know what we're working with. But it can be challenging to articulate those complex concepts to the layperson in the community. So. So I'm definitely going to give that a shot. That's such a great idea.
- [46:54]
Aziz Agaev
Exactly. I'll give you one more. As a business official, you're going to a board meeting and you're preparing a presentation. You upload your presentation and say, here's my audience. My audience are parents and board members. What might be the questions come up for this presentation, and then you prepare for those questions before you go in.
- [47:17]
John Brucato
That's awesome.
- [47:18]
Aziz Agaev
Like, these things are like, everyone can use it. Correct? Like, you don't need to be.
- [47:25]
John Brucato
It's not some secret that. Yeah, exactly.
- [47:28]
Aziz Agaev
You don't need to be tech savvy. You just upload your document and ask one simple question. That's it. I'll give you one more. Some of our districts upload their board meetings on YouTube and Gemini has an extension. You can say, at YouTube, add the link and say, summarize this meeting for me.
- [47:57]
John Brucato
Wow.
- [47:58]
Aziz Agaev
Take that summary. Give it to another chat where you train your board. Minute Templates, you know, board official, official minutes templates. And say change this summary to minutes and it will create minutes for you.
- [48:17]
John Brucato
Wow. I mean, that alone. I mean, some meetings go for hours. So imagine just being able to process that in, in a matter of minutes. That's incredible.
- [48:26]
Aziz Agaev
Yeah. Another, another use I would say is there's this app called Fireflies. That's what I use usually during the virtual meetings. I use Fathom to take notes and it records and it takes notes and then it summarizes the meeting. It all, it highlights all the questions asked. That you can just click and go to the question. And the Fireflies does that during the live meetings. You can just open it, put it on the desk. It listens to the meeting and then it says, this person said this, this person said that. And then it transcribes the whole meeting. Then it summarizes it, it gives you an overview, it gives you all the questions and it gives you all the tasks.
- [49:05]
John Brucato
All right, you know what I'm going to try is he's. So you've gone through a lot of different individual products. When I finish processing this audio, I'm going to give it to AI and say, can you pull out all of the products that Aziz mentioned so I can put them in the show notes of the episode? I'm sure it'll probably do that, right?
- [49:24]
Aziz Agaev
It will. Yeah, exactly. That's what AI is for. You just give a transcription and it will read that transcription and it will do exactly what you said.
- [49:34]
John Brucato
You're changing my life in real time, Aziz. I really appreciate it.
- [49:38]
Aziz Agaev
Yeah, no worries. One, one more thing before we go is some people don't know that this exists, but you can ask AI to create complex Excel and Google sheet formulas for you.
- [49:51]
John Brucato
Tell me more about that.
- [49:55]
Aziz Agaev
And I do get into the detailed demos of all of. I said during the training, basically what the important thing is to give exactly what you want, right. If you're specific enough what you're trying to do in Excel, it will give you exactly what you need. You just take it, paste it and it works. I'll give you a use case for that. You're calculating salaries. You have list of people in one sheet with their lane and step. And then you have another sheet that has the scale, right? Two dimensional scale with lanes and steps. You can say my column A does this, column B does this, column C has this information and all of that. Then column E, I want to calculate my salary on another sheet. I have the scale, it starts at B2 to J12. And you describe it exactly how you would see it. And then you say, write me a formula that would look into that scale and bring this correct salary based on lane and step.
- [51:06]
John Brucato
Wow.
- [51:07]
Aziz Agaev
And it'll create an index formula or vlookup formula, whatever the formula. And sometimes I do it and I don't know what it does the formula, but you can also read it explains the structure, it explains every single formula. Sometimes it just works, right? And sometimes you have long formulas that you want to shorten. I would just take the long formula, put it in ChatGPT and say, can I shorten it?
- [51:30]
John Brucato
Right. Right.
- [51:31]
Aziz Agaev
Is there a simpler way? And it sometimes spits out a simpler way. So even on a daily basis we use Excel, we use Google Sheet on a daily basis, we write formulas. Even if we use that with ChatGPT, that will change our life. And the difference between Google and ChatGPT is this. There is no examples online that relates to us. Trust me.
- [51:58]
John Brucato
No. And I can speak from experience because I will admit I still Google formulas every now and then. And it usually takes three or four different Reddit clicks or like, you know, something to get an idea of what they did and then kind of play with it to make it work for exactly what I'm doing. So what you're saying is you're taking your exact data and your exact thought process and it's tailoring a formula to exactly what you want. You're not reverse engineering some stranger, some online somewhere?
- [52:25]
Aziz Agaev
Exactly. That's the difference between AI and Google. Google gives an example of wide industries that are not schools.
- [52:34]
John Brucato
Right?
- [52:35]
Aziz Agaev
AI, you can tailor it to your needs. You just ask a question based on your data and it spits out what you need. Things like this, yes, it's little thing here and there, but that little thing here and there can save us tons of tons of time that we can then take our time and focus on a bigger picture. We always want to work on procedure manuals, we always want to work on, you know, five year plans for our offices, you know, for our team. But we never have time because we are always, you know, consumed with the daily little tasks. Right now it's time to give our little tasks to these little agents that we have and we can have tons of them and use our time for better things that can serve the kids.
- [53:30]
John Brucato
So what do you see as the next big AI innovation in school finance? Where do you see this heading in the next few years?
- [53:38]
Aziz Agaev
One thing I'm testing right now, and I'm thinking that that will be a huge change in the school industry. Is the operator. I don't know if you've seen the operator in action. It's ChatGPT operator. Basically you give it a command and then it opens a browser, it navigates to your browser and then it starts doing things in the browser while you watch. You don't have to watch, you can just go and then come back when it's done. But once this technology improves, imagine this. You feed hundred invoices and say enter these invoices to my ERP and it enters all the invoices to ERP and then you just come and approve.
- [54:27]
John Brucato
Wow.
- [54:28]
Aziz Agaev
And this is possible. Like right now, every school district does state reporting. And as we all know, state reporting is. We love them, right?
- [54:40]
John Brucato
Oh, absolutely. That's one thing where I have to put the brakes on AI. I can't take my state reporting away from me. That's the sacred category.
- [54:48]
Aziz Agaev
This takes time, tons of tons of time because there's a lot of manual entry. Now imagine an agent would take your data from Excel that you give it and goes to state reporting. Website goes next, next, next, finds the boxes, fills the boxes in and then goes next, next, next.
- [55:09]
John Brucato
Oh, that would be amazing.
- [55:11]
Aziz Agaev
And this can you. You can just say do not submit it it, but fill it up.
- [55:17]
John Brucato
Right?
- [55:18]
Aziz Agaev
Review is a lot easier than entry. That's how I look at it. So the operator part of this AI journey will drastically contribute to our day to day because we do a lot of things on our ERP systems for accounts payable for payroll. Imagine feeding 100 different timesheets and saying read these timesheets and enter it to a payroll system. And now I will just read and confirm.
- [55:47]
John Brucato
So is the operator function, is that something that's being developed right now by.
- [55:52]
Aziz Agaev
It is. It is available for pro users at this point for ChatGPT Pro users. And they were talking about having it available for plus users soon. I don't know when they will bring it to plus users, but it is, yeah, it's right there. And like another one is Google Sheets. If it's smart enough and if you train it well, you can now say, here's my Google sheet with different tabs. Can you create a dynamic formulas for me to do certain analysis? Right. It can go to Google Sheets, it can read your data and yeah, the possibilities are there. You know, it's just that we need to think of the use cases and try those use cases and we should fail during those use cases. We should fail again and again and again and then until we get very comfortable with it.
- [56:49]
John Brucato
So I'm Thinking about this operator function and, you know, in my own practical use, and I may need. I'm trying to refine my process. But, for instance, if I have my high school budget in Excel, each department is its own tab. And so what I do every year is I export each tab as its own sheet. So I can send it and then I compile it again. But I've come up with this ridiculous VB script to do it automatically for me, and it's never quite right. And I have to manually change some formulas. So the aioperator function sounds like this could be my golden ticket for making my life a little bit easier with kind of splitting these sheets apart and bringing them back together.
- [57:28]
Aziz Agaev
Yes, yes, it can. And actually, I do have a software for that particular thing. But on that aside, yes, AI can. You can just give it a spreadsheet. You can say, create me. And like, imagine this. You give a list of departments and a template, and then you can say, here's the template, here's the list. Create this template for every single department. Right. But even more, once you collect that information, you can ask, can you go in each department and see what's missing if there is any issues, if they're missing notes or if they're missing detail or if they're missing this, missing that. All of those things that we do on a daily basis now can be replaced with AI. And AI is coming to, I would say, a lot of software that we use, and we are also trying to bring it to our budgeting software, too, where you can actually talk to your data, you can ask questions to your data, and data will answer you. Right. So those things would be available now that you can ask who is missing the submission, who is working on the submission, who is missing the detail that they need to provide the. I mean, yeah, again, the questions are out there. Basically, the rule of thumb is if you think AI cannot do it, try it first.
- [58:58]
John Brucato
Right.
- [58:59]
Aziz Agaev
Because I'll give you one more. I was driving with my daughter and we were bored, and we said, let's just sing and have AI evaluate our singing. And my daughter sang. And AI said, in this part of the song, you need a higher pitch. In this part of the song, you need to change your voice to this. It evaluated. And then I said, let me try. I sang, and it said, maybe you should need another song, which means you can't sing.
- [59:34]
John Brucato
So it's at least polite about it. Right, Exactly.
- [59:36]
Aziz Agaev
Yeah. But things that, you know, it wasn't the thing that we knew that it does it, but we tried it and it did it. So if you don't know it does it, try it and it will probably do it.
- [59:50]
John Brucato
That's awesome. So as we wind down for any school of business officials that may remain skeptical about the use of AI, what piece of advice would you give them today?
- [60:00]
Aziz Agaev
I would say get out there and start using it. There is no other way of learning it. If you don't use it, the most important thing is to fail. If you don't fail, you won't be able to learn. And that's the if I had only one advice. It doesn't matter what they use. They can use Grok, they can use ChatGPT, they can use Gemini, whatever they want to use. Pick one and start interacting. That's where it starts.
- [60:34]
John Brucato
Well Aziz, thank you so much. You've gotten me re energized about using AI and I use it on a semi regular basis already and I really appreciate your time and I'm looking forward. Maybe I can pick your brain in another six or eight months and just see what has evolved from there and we can get some more tips and tricks for school business officials.
- [60:51]
Aziz Agaev
Sure. It changes every single day and that's why I have to work on a daily basis to make sure that I keep up. But after six months there may be things that we didn't even think that was possible.
- [61:06]
John Brucato
Great. Well, thank you for all of your work and coming on School Business Insider today.
- [61:10]
Aziz Agaev
Thank you for having me.
- [61:13]
John Brucato
Thank you for tuning in to School Business Insider. Make sure to check back each week for your favorite topics on school.