Second Drive Sales Wizard: The Ultimate Lead Generator

• 

  High Level Design Diagram of the Sales Agent

Inspiration

Buying a car is an emotional purchase for many people. They often prefer talking to a sales agent before making such a big decision. In some cases, customers might be on the fence, and a well-timed nudge from the agent can help them take the plunge. An AI agent with several sales tricks at its disposal can be persuasive in encouraging users to consider a purchase.

Other than that, Customers may have numerous questions about the cars available, and a human sales agent can't answer inquiries about every vehicle in the inventory. This is where an AI agent can be incredibly helpful.

What it does

The AI Sales agent chatbot helps customers with inquiries about used cars. The Agent’s goal is to convince customers to schedule a test drive for a car they're interested in.

The Agent utilizes the Gemini function calling feature to translate user queries into Postgres queries. If a customer expresses interest in a particular car, the chatbot guides them through the process of booking a test drive.

How we built it

It is built on Python and uses the LLM framework Langchain. It uses Gemini as the LLM model.

Challenges we ran into

It was challenging to convert the user's natural language to SQL queries. Several techniques were tried before reaching the optimal solution. Also, building the Agent was challenging too, especially the orchestration part making sure the data flows from one component to another.

Accomplishments that we're proud of

The final outcome that the agent is able to produce.

What we learned

What's next for Second Drive Sales Wizard: The Ultimate Lead Generator

Currently, the agent assists customers with inquiries and schedules test drives for cars customers are interested in.

Looking ahead, we plan to equip the agent with advanced sales knowledge. This will allow it to better understand customer requirements and provide more tailored responses. Additionally, recognizing that car buying can be an emotional experience, we aim to incorporate an emotional touch into the agent's replies.

Built With

• gemini
• langchain
• postgresql
• python
• streamlit