ChopCompass: How I built an AI-Powered Chatbot in 30 minutes to Discover the Best Dining Experiences

Introduction
On the bus to MEST campus today(2nd bus). I sat next to a colleague Andrews Kwesi Ankomahene and we had a light chat about a recently launched hackathon by AWS on devpost. The PartyRock Generative AI Hackathon by AWS: a challenge for all builders to get hands-on making generative AI-powered apps. In this challenge, we are required to use PartyRock, an Amazon Bedrock Playground as a fast and fun way to learn about Prompt Engineering and Foundational Models (FMs) to build a functional app with generative AI.
We are also expected to experiment with prompts to build an app based on one of four challenge categories or Remix the featured Side Quest app for more ways to compete!
The challenge categories are below:
- Interactive Learning Experience — virtual science lab, language lesson, etc;
- Creative Assistant — music composition assistance, story idea generator, etc;
- Experimental Entertainment — interactive story, Role Playing Games, etc;
- Freestyle — any project that doesn’t fall into one of the categories above.
You can either create a Party Rock app from scratch or Remix an existing one. I chose to do the latter being that it was already looking like what I envisioned.
Backstory and genesis!
Discovering the perfect dining experience can be a challenge but not anymore. As a firm believer in innovation and reinvention, I was inspired by the convenience and personalized recommendations offered by Google Maps. This inspiration led me on a journey to develop an AI-powered chatbot that would revolutionize the way people find the best restaurants.
Welcome to the story of ChopCompass, my simple chat bot solution that quickly helps you in discovering exceptional dining experiences based on your location, budget, cuisine preference, and meal you want to eat.
Finding Inspiration in Google Maps
The idea for ChopCompass was born out of the desire to create a user-friendly platform that could provide personalized recommendations for dining establishments. Seeing how Google Maps effortlessly provided directions and suggestions based on user preferences, I recognized the immense potential for a similar concept in the food industry. With this inspiration in mind, I wanted to develop an AI-powered chatbot that would take restaurant discovery to the next level.
Developing ChopCompass: A Simple Walk through
- I visited The Party Rock website as shown in the image below

2. I scrolled to the section (Try out featured apps built with PartyRock)

3. I clicked on the Good Eats App because I had something similar in mind and when I did; I was redirected to the page displayed below: This is because I already had created a Party Rock account. But if you don’t have an account it will prompt you to create one and you can use apple or Google SAML.

4. I clicked on the Remix button In the top Right corner

Immediately I did; it spun up my own instance of the Good Eats App.
5. From here on i added widgets; the first widget I added was the budget widget

6. After adding the widget; I proceeded to save the changes and published the app; subsequently you have to keep on pressing the release changes button after clicking save
7. One important feature I loved was the Show Configuration button on each of the widgets; It helps you choose the LLM you want to use and set the parameters like temperature and Top.

8. Finally once the app was ready I launched it.
What Next:
You can find Chop Compass here and test it out
Side Note: While developing ChopCompass, I love the fact that AWS party rock comes with about 7 open-source base LLMs and the UI of the site is leaning towards Neo-brutalism. I love it.
Testing and Iteration: Refining ChopCompass
Once the development was complete, it was time to put ChopCompass to the test.
User feedback plays a vital role in refining the functionality of products and ensuring their effectiveness. I shared a link to the remix app in our group of fellow Global entrepreneurs so they could get to test it out and use it extensively.
I got some exciting feedback and also got two people to remix my app and go ahead to create their own apps. The feedback they gave also enabled me to identify areas for improvement and address any usability issues.
Reflection and Future Enhancements
Creating ChopCompass was a swift and exciting process that took only 30 minutes. I utilized the no-code platform, AWS Party Rock, to build the app. I started by remixing an existing app, adding more widgets, and making a few edits to customize it to my needs. Within a short time, I had successfully spun up ChopCompass on the AWS Party Rock platform.
The development process was straightforward, and I encountered no significant challenges along the way. ChopCompass was built with a focus on user-centric design, leveraging the convenience of AI-powered chatbots and the personalized recommendations of Google Maps. The app’s core features include the ability to provide tailored dining suggestions based on user preferences such as location, budget, cuisine type, and meal type.
My experience building ChopCompass has been both educational and rewarding. I learned the importance of prioritizing user experience and the potential of AI in transforming everyday experiences. I was able to efficiently create a unique and functional app in a short amount of time using a no-code platform.
Moving forward, I plan to enhance ChopCompass by incorporating advanced machine learning algorithms to improve the chatbot’s recommendations and expand its reach.
I believe that ChopCompass has the potential to revolutionize the way people discover dining experiences and look forward to sharing my journey with you.
TL:DR
- ChopCompass was developed in a hackathon.
- ChopCompass was built entirely on a no-code platform called AWS party rock
- It utilises the Claude LLM
- The chatbot’s recommendations are based on a combination of user cuisine preferences, location, and budget.
- The app allows users to scout for and find good diners and restaurants they can visit and patronise with friends.