I have always had a profound interest in the human mind. I have taken courses and paid for learning how to influence the human mind and how we can better learn to control it. And a few years back, my friend after recieving a pacemaker started going through bouts of depression and paranoia.

She couldn't afford therapy as it was so expensive. A couple of us friends pooled in together to help her manage, however, she wasn't willing to trust and open up to anyone. After a lot of sleepless nights for all of us, we finally convinced her to seek therapy, where she was partially diagnosed with PPD. But as people suffering from PPD don't really go to therapy and research on it was limited, there werent exact techniques that could be used to help her out.

Luckily, around the same time research started to come out about how a few psychiatrists from Stanford University had used bots to help people suffering from depression and other mental health related issues. Research showed that people who otherwise might be shy in opening up, found it easier to talk with machines. I also did a lot of study on individuals who suffer from paranoid personality disorder (PPD) and learned, that PPD often is a precursor towards depression, increased anxiety and schizophrenia.

As part of my master's thesis, I worked towards modelling the personality of individuals who suffer from PPD using AI avatars, natural language processing and data mining techniques with the goal that this may help psychiatrists and mental health professionals to better understand PPD and also other mental health issues, and being able to improve therapy techniques in order to better help their clients.

Furthermore, after diving deeper into this topic, I found out that mental health professionals are often overworked, and in third world countries there is 1 psychiatrist for 100,000 people according to the WHO.

Pre-Covid 19, depression was being expected to be the number one disease by 2030, and due to the isolation and uncertain times we all face now, this expectation is likely going to increase exponentially. So, we now more than ever, need new ways to help both the professionals with the skills to help and the people who could benefit from them.

Digital assistants can help, with more and more research coming out to explain their benefits and uses. However, they can be expensive to build and not every mental health professional has the resources or know how. And so in order solve this problem the idea of EVAAI was concieved.

What it does

EVAAI has been designed for one purpose that is to help mental health professionals create AI digital assistants without having any need to code. Our aim is to make the process of training and scripting digital assistants as simple as typing a document in MS word. These assistants can help them provide basic counselling to their clients, help them meditate, measure their moods, help follow a routine, manage conversational data of patients easily for when they are not available.

These assistants are available 24/7 and since they replicate the language patterns of the professionals that built them, the clients feel a sense of rapport, similar to the one they feel with a therapist they are comfortable with. As opposed to random volunteers, helping them out. This helps ease the burden on the mental health professionals and personalises the experience for the clients. The assistants also set up appointments with clients for providing in depth online and in person therapy.

EVAAI is also a community that brings together mental health professionals from all over the world to easily share, discuss and come up with novel research with each other in order to help all of us in need.

How I built it

Our tech stack includes tomcat, java and javascript for Natural Language Processing, postgresql for database and python and tensorflow for deep learning analytics.

Challenges I ran into

This is a complex problem to solve. And we've had to conduct many private beta tests, surveys, and interviews. We'd also have had to test all the existing technologies out there, so we could improve and find the right mix and match for our solution. We will be improving on the backend engine during this hackathon.

Accomplishments that I'm proud of

The fact that I had a specialised interest that allowed me to able to look at this problem from a unique angle, along with finding the right people to execute this idea, the diligent hours of research in better understanding the people we are suppose to serve, and intelligently using our limited resources to bring this idea so close to a proper launch, is something me and the people who have helped me in this journey are exceptionally proud of.

Furthermore, the excellent response that we have recieved from our target market with 30 mental health professionals, who collectively deal with hundreds of clients, willing to develop their assistants and join in the community from all over the world, has really motivated us to keep pursuing our idea and officially launch.

What I learned

Understanding the pains of the people you serve is the most important thing.

What's next for EVAAI

To partner with more institutions to help them reach out to more people, who otherwise might be shy and unable to afford therapy.

Also help more mental health professionals to become more self sufficient which allows them to help more people while charging a competitive rate that their clients can afford. Thus, helping both sides.

Find funding in order to hire more capable people and keep the original team in place so we can streamline the process of development and help as many people as quickly we can, both during and after the Covid-19 crisis.



figma, java, javascript, postgresql, python, tensorflow, tomcat

Devpost Software Identifier