Dashy Assistant

• 

  Dashy Assistant

• 

  Dashy Data

Inspiration

You can only do so much on a small mobile screen. No matter how responsive or mobile friendly the app is, viewing complex data on a smartphone is not an efficient endeavor.

What it does

It's a voice assistant for proptech data. In this prototype, for multi-family property management data.

How I built it

Using AWS services powered by things like automatic speech recognition (ASR) for converting speech to text, and natural language understanding (NLU).

Challenges I ran into

Building a voice interface without a direct path to real data is challenging. Just setting up the test data and structure was time consuming.

Accomplishments that I'm proud of

Even though I've never built a voice interface before, the result was incredibly promising and quite exciting. This is just scratching the surface of what's possible.

What I learned

Conversational interfaces, like the one that I created can be very powerful. They are not however magical and do require a lot of effort to work well, especially when dealing with domain specific topics, like real estate and property management. Asking for the weather is one thing, requesting budgeted collected occupancy % is a completely different animal.

What's next for Dashy

I'd love to expand the Dashy Assistant to include visualizations. Even though voice is incredibly efficient for making requests, it's sometimes lacking for receiving responses. So, for example, when requesting new prospect data, it would be so valuable to not only include a direct voice response, but also an on-screen visualization that could include things like prospect trend charts.

Built With

• amazon-web-services