Even on best UI designs it takes a lot of click to set your filters and get the search go! Rebecca is here to help you search a house just by describing it (via voice on applicable platforms)

What it does

Rebecca can help you find property (now is connected to our Malaysia API) as simple as talking to your friend. She is a smooth talker and does not let you feel talking to a emotionless bot! She can talk to you about anything; even telling a joke or talk about Obama!!

How we built it

We used AIML for the static chit chat and machine learning via API.AI interface for more complicated conversations. The extracted information will later sent to Lambda function which based on the Intent can respond with chit-chat, house listing search or even weather forecast!!

Challenges we ran into

The AIML itself is a challenge as it require proper setup and scripting to get a human-like response. The machine learning part is also challenging when it comes to providing the training sets as wrong info will teach wrong model!

Accomplishments that we're proud of

Rebecca is literally the first Real Estate Bot in South East Asia. Rebecca is far far more intelligent than any Real Estate Bot. Rebecca AIML scripts are almost as perfect when it comes to casual conversation.

What we learned

Serverless computing rocks! No worry about scalability or performance.

What's next for Rebecca

Deep Learning. Going to improve the NLP as it is the key factor in user acceptance.

Built With

Share this project: