“Business Buyer Software”:
This is a tool to help potential business buyers make more informed decisions when trying to buy a business.
https://www.loom.com/share/963bee99987446a9be07314403421d37?sid=cb898880-30e4-445b-a2d9-efb1d7603c29
fincalc.jeremiahwoldabezgi.site
"Should I Invest into this business?" "Is this a good amount of annual income fora [Insert business type] type of business?" "How much profit would I make after loan payments after 5 years?" "Is my interest rate on my loan high for a SBA(7a) loan"
Name: Jeremiah Woldabezgi
Vandy email: [email protected]
VUNetID: Woldabjs
Role: Sole Project Developer
There is so much analysis that goes into decisions regarding the purchase of a business. Large institutional business buyers hire analysts who’s sole role is to help the decision makers make better decisions by analyzing data and demonstrating that data to the decision makers in consummable way. Regular everyday business buyers can’t afford to hire analysts for this role.
Business Buyer Suite is a software tool that helps users make more informed and more well thought out decisions on if they should buy a business or not by helping the owner project out into the future and estimate what will happen to the business given different scenarios using financial models and projections.
It can understand user requirements through natural language conversations and guide them through the acquisition process. This ensures users have all the necessary information and insights to make well-informed, intelligent decisions quickly. It also allows them to focus on other important tasks, such as speaking with business owners to source potential deals and securing financing for promising opportunities.
Develop an AI-driven software that assists users in generating financial models and projections after natural language conversations with the program.
- The software will act as an assistant, providing information, scenario analyses, ROI analyses, and more.
- Through user-friendly interactions similar to the interactions between a decision maker and an analyst (natural language), the user can get most of the benefits of hiring a full time analyst for a tiny fraction of the financial costs and an even smaller fraction of the time that it takes to work with an analyst.
- Requirement Gathering and Analysis - Understand all the functionalities and features the software needs. Understand the pain points of business buyers to see what problems relating to analysis affect them the most.
- Development of Core AI Engine - This will include natural language processing capabilities, financial model generation, etc.
- Interface Design and Development - Begin with a simple interface probably just a textual conversation where the tool could just return spreadsheets. But I will need to try out different potential interfaces that could be used for this tool to find the most effective way for a user to interact with a tool like this.
- Integration of the AI Engine with the Interface - Ensure a seamless interaction between the user and the software.
- Testing and Refinement - Test the software, refine based on feedback.
The tool needs to provide accurate and understandable financial models.
The tool needs to Educate users on the significance of the generated financial models and how they can interpret them, ensuring they can actually use them to make faster and more informed decisions.
Offer a way to see what would happen in different scenarios and given different input variables so the user can see how different variables can affect the outcome (focusing on how everything affects the potential “seller's discretionary income” especially)
Enhance the credibility and negotiation power of business buyers in front of potential investors and banks by providing them thorough analysis of the potential investment (which investors/banks really like)
Ensure the software is user-friendly, and can iteratively guide and educate users all the way from initial thoughts and inputs about the business all the way to their their a final decision. (For example asking for more information when more information can help the program generate better financial models vs just making a poor analysis using only whatever information the user has already provided)
Measurement: Evaluate the correctness of the financial models generated by the software
- A Grade: Less than 5% error rate and the model was able to catch its own errors.
- B Grade: 5-15% error rate and the model was able to catch its own errors.
- C Grade: 15-25% error rate and the model wasn’t able to catch its own errors.
- D Grade: 25-40% error rate and the model wasn’t able to catch its own errors.
- F Grade: Over 40% error rate and the model wasn’t able to catch its own errors.
Measurement: Collect user feedback on the naturalness of the interaction with the tool and how well it explained generated models.
- A Grade: Most users found the interaction very natural and models very comprehensible.
- B Grade: Most users found the interaction natural and models comprehensible.
- C Grade: Most users felt the interaction was somewhat natural and models were somewhat comprehensible.
- D Grade: Most users felt the interaction was not very natural and had difficulty understanding models.
- F Grade: Most users found the interaction unnatural and models hard to understand.
Measurement: Assess if the tool ensured it had all the necessary data before generating financial models.
- A Grade: In over 95% of cases, the tool verified it had complete data before generating models.
- B Grade: In 85-95% of cases, the tool verified it had complete data before generating models.
- C Grade: In 75-85% of cases, the tool verified it had complete data before generating models.
- D Grade: In 65-75% of cases, the tool verified it had complete data before generating models.
- F Grade: In less than 65% of cases, the tool verified it had complete data before generating models.
Address each of the goals, and assess each of the metrics. Include a statement on each on what you achieved or did not achieve, give support for your assessments.
I changed the direction of my project a little bit while working on it because I realized that it could be a lot more useful if I focused on it being a calculator and integrating Chat GPT by adding the qualitative analysis of my data from their ai large language model.
For goal one I was supposed to make sure my tool provides accurate and understandable financial models and I think I met that really well given the scope of this project
For goal two, I was supposed to users on the significance of generated models and how they can interpret them, and ensuring that they can actually use them properly to make informed decisions. I think this Was done really well when I integrated Chachi to help talk to the user about the data and the financial situation of the business.
For go three it's supposed to offer a way to see what happens in different scenarios and give different input variables so the user can see how different things affect the outcome. Turning this into a calculator with many inputs that can each be changed to generate a different set of outputs. Does this exactly I think this was completed well.
For four I was supposed to enhance the credibility negotiation, power of business buyers in front of investors and banks, by providing them analysis of potential investment. I think this is also done well, because the tool gives you clean and clear financial metrics that are very useful in determining the profitability and investment viability of the business and Chat GPT will explain everything about the data extremely well for the user, and can give the user very intelligent things to say to investors better in the business.
For goal five I was supposed to make sure the software is user-friendly and can guide and educate users all the way from initial thoughts into their decision and I think this is done well too because it's a simple calculator with simple inputs that are clear and you can ask questions to the model to help Guide you along the way all the way to the position where you can make a decision about the potential business purchase.
What do you take away from the project? Has this changed how you understand AI? Does and how does this affect future plans for learning, work, or otherwise?
I learned so so much during this project. ChatGPT taught me CSS and JavaScript very quickly, I'm actually very impressed with how useful chatgpt actually is in being a teacher, when you are trying to learn by doing alongside a project versus learning everything out of a textbook or from a lecture. Also, I really could sense how ChatGpt literally just tells you what it thinks users often say in different situations versus it actually thinking sometimes that becomes very clear.
I had to learn everything about web development, including javascript, HTML, and CSS in a short period of time. I built the site from scratch with the help of ChatGPT guiding me while I was developing it. I learned extremely fast because I had chat to PT the entire time helping me along the way answering some very specific questions telling me every detail about how the development works and helping me program and debug the software I used to make this app
Do you plan on continuing the project? What will you do with what you've learning?
I definitely plan on continuing this project a lot I'm gonna keep adding more and more features to the app, the help of chat, including more charts graphs, and interactive tools. I also, plan on adding a feature where a user can talk about their potential business using words, and then Chat GPT will extract the relevant numbers from what the user says, and inputs statement of the calculator automatically.
You can run the program live using my hosted website, however if you want to run it locally on your computer, you can use these instructions to do so:
Open a terminal in the folder you cloned this repository in
-
First install Node.js onto your computer, you can do this at this link: https://nodejs.org/
-
Then clone the repository and place it in a convenient folder on your computer
-
Then open the .env.example file in a basic text or code editor and replace the 'insert_your_api_key_here' with your API key from OpenAI
and then remove the .example from the .env.example filename until the file is now just called .env
-
For example:
-
Before:
-
After:
-
-
Next, open a terminal in the cloned repository's folder
-
Run this command in your terminal (while in this folder) -- 'sudo node server.js'
-
Open http://localhost:3000 in your browser to see the webapp running locally on your computer!