Business Development, Customer Development and general sales activities are time consuming and often do not provide enough return on both time and monay investments. Moreover, consumers wants to buy or get their service done as soon as and as fast as possible, not regarding time of the day.

What it does

That's why we have created a 3D Virtual Business Assistant for Web & AR, which is much more engaging than a regular text chatbot and allows customers to learn more about vRational brand and schedule a project consultation with one of our consultants. Each appointment is validated with our MakePlan Business Calendar and confirmation is sent to the customer via email.

Customer can communicate with our chatbot naturally using voice or by selecting one of the propositions displayed in response.

How we built it

Application core was built with Amazon Sumerian and provided Host component, Amazon Lex with a custom built AWS Lambda function wrote in Python. We also use text-to-speech AWS Polly service to allow our Virtual Assistant to respond with voice.

It works both on web and as an AR application (in a separate AR-enabled scene, which have to be wrapped in provided ARCore starter Android project.)

Challenges we ran into

It was challenging to design UX that which works on mobile and desktop web, and in AR application. We had to write a lot of custom JS scripts to unify expected Host behaviour.

Accomplishments that we're proud of

We are proud that we were able to create User Interface, which works both on mobile, desktop web and AR. User can communicate seamlessly using Voice, Text and clickable prompts.

What we learned

We were able to learn and utilize to the maximum Amazon Sumerian framework with custom JS scripts and CSS classes. We also discovered how quickly develop endpoints with AWS SAM to enhance standard Lex chatbots. We will be able to even more engaging and interactive chatbots in the future.

What's next for vRational - Virtual Business Assistant

We were able to create a general Core Asset Package, which we can reuse in other similar projects and various Lambda functions, so we might consider creating a product with it. We will still work more in Amazon Sumerian and other AWS services, as we've just started exploring their vast possibilities.

Built With

Share this project: