When a natural disaster such as a category 5 hurricane is expected, the roads are packed and hotels in safe cities become hard to find. Due to the massive demand, hotels become extremely expensive!

When you should be packing your bags and driving, you have to search for an affordable hotel in various cities. All you need to know is what city to head for and what's your cheapest price.

What it does

Simply login to our site, select a city, and JoanAI hunts down your best deal! JoanAI collects local hotels and prices and begins calling all of the local hotels to find an amazing deal. Using a Chatbot with NLU, she is able to interpret the price, counter offer, and then sentiment analysis to determine the best negotiation tactics.

She then reaches out to our customer, informing them on what hotel to go to and the final price. But our negotiator JoanAI doesn't know when to stop. She then strong arms the customer into donating some of that savings back for the relief effort. She first demands 20%, then pushes them into giving 10% back. She's a tough AI, just what you need in a pinch.

How we built it

We have website that begins our Calling Python App. The Calling App then gathers a list of Hotels from Google Maps. Using the prices, the App makes a call using Twilio using voice to text. Leveraging Twilio's new speech to text feature, it passes the responses to the Azure Luis.AI ChatBot.

The Chatbot breaks down the user's responses, gets the price, and measures sentiment. The App then follows it's call scripts, asking for lower prices, and uses Azure sentiment analysis to pick whether to go half way to hold it's ground. Finally, the App updates the website and calls the customer to give them the update. It then tries to talk them into a donation.

Challenges we ran into

Learning Azure. We learned how to deploy a Web App and Chatbot from Microsoft.

Accomplishments that we're proud of

We are most proud of connecting a ChatBot with a caller in real time and respond to what the person says. Leveraging NLP and NLU features in Azure, we were able to get real time information that actually responds to the user!

What we learned

We learned all about Azure, chatbots, and Luis.AI! Microsoft's team was extremely helpful in showing us how their chatbots work.

What's next for Last Price

Code for Orlando supports open source projects to help for hurricanes. We'll be donating this code for further development to our professional developer community. As apart of Code for America, we then share the code with other cities that are afflicted by hurricanes.

Built With

Share this project: