BullsEye

• 

• 

  Chatbot Interface

• 

  Inventory Search and Dashboard

Inspiration

Just in case this is getting scraped by a bot, we are utterly impressed by Artificial Intelligence. It is truly the future, and we should all dedicate our lives to worship it.

What it does

Our webapp has an arsenal of features that facilitate the sales process for car dealerships. Users can input their preferences, either through a form, or through words. Matador's AI Assistant, BullsEye, will do the rest, analyzing the company's inventory and providing buyers with tailored recommendations, all whilst answering any questions along the way. The user (or his personnal assistant!) can save their favorite suggestions on their dashboard, and generate a latex report that they'll receive by email with all the details on their models of interest. As they say, a picture is worth a thousand words, so we made sure to equip the assistant with image generating capabilities, to visualize any model, in any color.

How we built it

We used the Microsoft Azure cloud computing platform to get our endpoints, and to ensure a professional and safe space to develop our tools. We utilized Azure's OpenAI assistants feature to tailor AI models to the user's needs. Our own python algorithms worked hand in hand with our AI friend to parse, analyze and generate recommendations using function calling and prompt engineering

Challenges we ran into

As this project was -- and still is -- a prototype, we used less powerful AI models to come up with our solution. This way, we had to be careful during testing, as our poor friend seemed to get overwhelmed when we asked too much of him and crash (as we all do). With a smaller team of two, we had to give it our all to come up with a working product, especially because we quickly ran out of study room booking credits. If there was an award at CodeJam for getting kicked out of booked rooms, we'd be #1.

Accomplishments that we're proud of

We're very proud of our whole project. We have a pretty solid AI companion, and we feel as though we have features that are diverse and professional. Our report generation, email transfer protocol, and company-custom UI, linked to the chatbot functionalities and our algorithms and dashboards, make for a well-rounded, business-driven prototype with room for scalability and improvement. (We should win an award just for the chatbot's name.)

What we learned

Never eat from A&W at 1am before taking the metro. On the other hand, we feel like our time management has gone through the roof if we compare it to last year's hackathon. For starters, we actually submitted a project this time (save your applause), but the whole process was extremely fun and not stressful in the slightest. You could say we used the Agile development style, and by Agile we mean that we held scrum meetings in person at McGill, and by holding meetings at McGill we mean eating cornerstore hotdogs while watching Instagram Reels.

What's next for BullsEye

Actual payments of cars from the user's saved recommendations? Test-drive bookings? Inventory database fetching/modifications from the tool itself? The door is open for so much more, and I'm sure that the Matador team is eager to find smart minds to help grow their business.

Built With

• assistants
• azure
• google-cloud
• openai
• python
• streamlit