BizRizz AI

We’ll make your business Rizzz and rise up!

Comment

Inspiration

Coming from a family having deep involvement in the restaurant and hospitality industry, I was always surrounded by it. Growing up in this environment made me naturally analyze everything around me. So many things could have been done differently, and many resources could have been saved. How can profits be increased?Well, I know, I know... consultancies exist, but a restaurant consultancy fee can vary between $40 - $120 per hour, with daily rates ranging from $250 to $1,000. Now, that's kind of expensive for small businesses. I thought, why not leverage my skills in AI and deploy an AI application that can actually end up being your free consultant for business? That's when I came up with BizRizz... lol -> ("We’ll make your business Rizzz and rise up!") that's the main motive: helping small restaurant and cafe owners actually boost their businesses. This will help them grow and also help the country's economy boost because the restaurant and foodservice industry is one of the largest in the United States, both in terms of its economic impact and the number of people employed—1 in 10 people work in the industry, which generates 2.5% of the United States’ GDP. (escoffier.edu)

What it does

The application is currently in the development stage. It takes user details from business owners like business name, location, and kind of restaurant. Then you ask questions regarding your business and, voila, it will return advice and suggestions by monitoring your competitors, checking your rating, customer feedback, and also by analyzing the basic details given by the user in their prompt.

BizRizz will then generate a report collating all this information and providing business advice.

Simple execution but complicated backend.

How I built it

So the base model for the project is the Gemini API by Google. It's free and works fine. I'm then using the Google Places API to fetch reviews of places and get an idea about the competitors. I review the competitors' ratings and analyze the total number of voters to check how well-known they are, because a restaurant with one rating of 5 vs. a restaurant with hundreds of reviews with a rating of 4 can't be compared easily. Along with the real reviews, I am identifying positive and negative reviews and implementing some analysis on that too. I am also utilizing the Paychex API for employee data, which will help me provide employee insights, but unfortunately, the API key is giving me some issues. I’ve incorporated it, but it gives no results really. Utilizing all this and doing prompt engineering, I built BizRizz.

Challenges I ran into

After I was done with my backend, I wanted to integrate my frontend. I had my entire backend utilizing FastAPI. I thought of using React for the frontend, but I am new to frontend development. I kind of figured the part out, but React wasn't working for some reason as it was not letting me install my Tailwind CSS. After going through a lot of documentation and videos, I gave up and used Flask for my backend instead and used HTML, CSS, and JS. I know it’s not the best choice, but it works for now.

Accomplishments that I am proud of

I used to always work on backend, but it's the first time I was able to develop the frontend. I managed to fully publish an actual working application. It’s in the development stage, so no hate lol!

What I learned

Tech skill-wise, I actually learned React, and I also learned some things like deploying tasks through applications, how to mix and merge everything together, and deploy it into a single product.

Non-tech wise, I learnt that I should sleep more the day before any hackathon lol.

What's next for BizRizz AI

As mentioned earlier, this is just a consultancy product that is currently working on small features and attributes. I want to deploy a RAG-based implementation where I can get owners' opinions, employees' opinions, and also a human consultant advisory to add more depth. Along with it, I aim to work on reinforcement learning by making a reward-based system to get better results. I want to add some new features that I still need to think about. I dream for this to be a fully scalable product, so I plan on forming a team to work on this. Because making this completely may seem like a one-man job, but it’s not. Also, instead of using Places API for reviews, I want to apply web scraping and NLP-based approaches to analyze the reviews and get positive and negative sentiments. The prompt engineering part is also something that needs improvement. Further more this can be made into a full based system by integrating computer vision models to get work insight on how the employees work and how the resources can be saved. That would probably make it more like a manager but It’s something that I look forward to as an actual AIaaS.

Share this project:

Updates

Submission history