For eSports psychologists our conversational bot provides their gaming team clients with preliminary self-help until the psychologist takes over for live consultation.

Inspiration

One of our team members is an aspiring professional gamer, already among North America's better players of Overwatch, a team based first person shooter favored by millions including Elon Musk and Sam Altman

When debugging mental challenges affecting gameplay , he discovered that there are professionals with deep expertise in helping gamers and their teams better cope with issues such as choking in high-pressure situations. But their services are not yet readily available on Discord, the most popular messaging platform for gamers.

Competitive computer gaming is a rapidly growing industry. Just as other professional sports, there is a sub-industry of eSports psychologists to help players and teams reach peak performance. Many players and teams are willing to make considerable investment in improving their gameplay.

Dream Team

Our team consists of:

  • Semi-professional competitive gamer
  • Biz person with degree in Psychology
  • Experienced software developer and entrepreneur

How we built it

Our bot targets the Discord messaging platform popular among gamers.

The bot builds on ideas from the classical ELIZA chatterbot from MIT AI Lab, using pattern matching and substitution to create an engaging psychoanalytic-like discourse. We customized and extended the patterns to cover domain specific issues of concern for gamers and gaming teams, rather than rely on a third party generic intent analytics API.

We implemented our own Dashbot SDK in the Python programming language, passing on to the Dashbot REST api the text of both incoming and outgoing messages. Our gaming specific NLP engine determines the 'intent' of messages, which we post to Dashbot using the REST api to improve the analytics on their dashboard.

Accomplishments that we're proud of

Chatbots often struggle to go beyond a canned dialog sequence. We designed Psycho to be flexible and resilient, able to make reasonable responses to a wide range of questions and concerns within its domain. We can use the analytics from Dashbot to further improve its responses, by learning which incoming messages are not handled well.

Built With

Share this project:

Updates