Inspiration

Though authorizing documents has reduced down to a click, that's still a hazzle for select elderly and differently abled. With the desire to make things simpler for these folks and also to take e-signatures to the next step, we bring to you SaySign - A 2018 new way to get things done.

What it does

What's better than signing a document... well electronic signatures are (said DocuSign in 2008). What's better than e-signatures that demand a click you ask? A completely hands-free way maybe...

SaySign works right on your mobile ! But wait... there's more! You don't need to download another app to let SaySign be of service to you. SaySign is a bot that interacts with you over your friendly neighborhood Google Assistant!!

SaySign brings document authentication to your phone, in a seamless way and theoretically is as robust as Google.

You can now ask your favorite voice assistant, the documents you have received, need to sign ASAP or even get a concise summary of your document. Did we mention you can sign and mail your insurance documents, transcripts, etc. right from your phone! Yes! You can Sign it just by Saying it !

How we built it

We built our bot using dialog-flow by Google. It functions with the help of Docusign API which gives it its skillz with the docz.

We used Heroku as our backend and our web-server was made using Python backed by Flask API.

We've used other APIs including Selenium, gensim and pypdf2 to aid us in our conquests.

Challenges we ran into

It was great working with all the API and learning and integrating everything over the past two days. We definitely encountered what seemed like a dead-end, more than once. But since we are a team of friends who didn't give up on each other and we managed to pull through.

A couple of our hurdles:

  • No api interface to add a signature to pdf
  • Firebase options being limited with the free account

Accomplishments that we're proud of

We tackled quite a few technologies which we weren't exactly familiar with. We didn't back down and didn't let sleep get to us :) We were able to make a working POC which will hopefully turn into something which many deem useful

What we learned

We have not worked on DialogFlow before but all of us are well aware of the concept of Natural Language Understanding. This was a new take on things and was fun learning and working with. Definitely a skill we are proud to have now.

We were initially exploring speaker diarization which made us explore a lot many math models, which was cool.

What's next for SaySign

We hope to use speaker diarization and speaker identification models, to add a robust layer of security, to standardize SaySign. SaySign could (machine) learn more and incorporate topic modeling for more accessibility.

Built With

Share this project:

Updates