‎BenzBot

An example of what chatting with the bot can look like

Inspiration

Our chatbot, designed to recommend the perfect car based on personal information, was inspired by the growing trend of personalization within technology. We aimed to enhance user satisfaction by creating a tool that not only provides tailored recommendations but also offers convenience and efficiency in the often time-consuming process of car shopping.

What it does

Leveraging data-driven decision-making, our chatbot delivers accurate suggestions that align with individual preferences and needs, addressing the complexity and variety in today’s automotive market. By simplifying the decision-making process and improving overall customer experience, we help users navigate their options more effectively, fostering a deeper sense of trust and loyalty

How we built it

To create our bot, we made use of Salesforce’s platform. First, we created customer profiles with a range of attributes relevant for deciding what car best suits them. With these attributes in mind, we used ChatGPT to synthesize data on various EV and Hybrid Mercedes-Benz cars. Then, we engineered a prompt and made use of more Salesforce tools that enabled us to insert customer data as placeholders within the prompt. This final prompt would then be sent to ChatGPT for it to make its vehicle recommendation.

Challenges we ran into

Lack of experience with AI: We are two bachelor students TU Munich studying Computer Science and Information Systems in our second semester. We have little to no experience with AI, which made this task particularly difficult for us. Thankfully, the salesforce platform proved to be an effective solution to this problem.
Overwhelming range of functionality: At the same time, the overwhelming range of functionality of the Salesforce platform was a difficulty in itself. With support from Salesforce experts, we got to know the relevant tools and were quickly able to create a functional prototype.
Engineering a reliable prompt: With this in mind, the last and arguably most important challenge was engineering a reliable prompt that always manages to create a personalized recommendation that resonates with the customer.

Accomplishments that we're proud of

Naturally, we are very proud of the fact that, starting with practically no experience with AI and chatbots, we were able to create a functional prototype in such a short period of time.

What we learned

Not only did we learn how to use the Salesforce platform, leveraging AI to create a personalized experience for customers, but we also learned how to create better prompts that lead to the desired output reliably.

What's next for BenzBot

In the near future, we are planning to create a fully interactive chatbot that can be integrated into the Mercedes website and be used by customers to promptly receive a vehicle recommendation based on their data and preferences.