A lot of phone calls received by the customers are repetitive. By outsourcing these questions to our salesbot, companies can lower the cost of manual labor significantly.

What it does

The salesbot answers questions based on a database provided to it. It’s able to update pricing by predicting sales by season using features such as brand,time of year, etc to apply automatic discounts to potentially maximize sales.

How we built it

We use neural networks to detect seasonal changes and adjust the price of sales, and based on a database generated by the neural network, then using the database in union with the openai question and answer knowledge base, we are able to provide a complete customer experience bot. we are currently using dummy data we created , but it is applicable to real data.

Challenges we ran into

We don’t know OpenAI and Jsonl prior to HackAI, so we have spent much of our time figuring out how OpenAI would behave based on the database it uses and how to correctly format data for OpenAI access.

Accomplishments that we're proud of

We are proud of the database we can generate using neural networks and how we integrate into our salebot build using OpenAI

What we learned

We learn alot about OpenAI and how it can be leveraged for many different task with minimal amounts of data specification.

What's next for SalesBot

If times permit, we would like to integrate salebot into a live database and provide a GUI for it, so the customer can have a more visually pleasant interaction with our bot.

Who are We

Arjun Balasubramanian Aaron Smith Jian Liang Tarun Paul Bethepu

Built With

Share this project: