Discourse

Speech recognition tool for supervisors and police officers working with sexual assault victims to track training, measure growth, and compare practical skills with their real-world experience.
Project Brief
Role
Designer
Duration
Nov 2020 - Jan 2021
My Contribution
User Experience
Interface Design
Research
Prototyping
Collaborators
Amanda Yeoh
Della Sigrest
Tara Janes
Background
This was a project created for IxDA's Student Design Charette 2021. The brief was on achieving greater collective wellbeing through the power of our individual data.

Taking this broad brief, our team initially discussed interests and after exploring topics like city systems, urban environments, palliative care. We finally converged around the problem of sexual assaults going unreported and its affects on the community.
The Challenge

Unreported cases lead to unresolved crimes and increasing dangers as perpetuators remain free

01
Most sexual assaults go unreported leading to unresolved crimes.
02
Insensitive reporting process without metrics to improve interactions
03
How to make individuals feel confident that sharing data will support their needs?
Research
Most cases are unreported because of the lack of trust and doubts on how a case will be handled.
After talking to advocates, SANE nurses, and law enforcement officers, we found that a majority of survivors or victims of sexual assault often don’t report their cases because of the complex system and uncertainty of whether they will get justice by the end of it.
The Existing Journey of the Victim / Survivor and the Police Officer / First Responder
Ideation
Analogous research
To get a fresh perspective on our problem, analogous research was used to identify other industries and different contexts where we could find similar activities, behaviors, and emotions.

We looked into VR training platforms, real-time feedback, voice user interfaces, and assistant bots. Following this, each team member individually ideated on how these can be used in our context using Crazy 8s.
Scoping down and validating ideas
There are numerous complex issues around sexual assaults and to make sure we reach a usable and valuable solution, we wanted to scope down our project.

Based on feedback from experts in law enforcement and voice recognition technology, we decided that voice recognition integrated with a trauma-informative tool would be the tool most comfortable to the user while making the most impact.
The Solution
Introducing Discourse
Discourse is a speech recognition tool for supervisors and police officers working with sexual assault victims to track training, measure growth, and compare practical skills with their real-world experience.
Roleplay Training
Officers can participate in role play based training, where Discourse analyzes the language and the context of the conversation to provide real-time suggestions. Non-verbal cues like pauses and change in tones are also identified to give live insights on how the victim is feeling.

This helps investigation officers easily understand when they need to explain their questions better or simply give more time to the victim to reply.
Feedback & Reports
Officers can also import voice recordings of conversations from actual cases, using which Discourse can generate a report to see how they compare to their trainings.

Based on the performance on actual cases, officers can train themselves in specific skills and participate in more role plays.
The data angle
Vocal narratives are the touchstone of the interactions between our users
Speech data is collected and analyzed for their sentiments from a trauma-informed perspective. Non-verbal cues are the most promising features in providing officers with insights about how to say something rather than what to say.

Having insights from Discourse would change officers’ interactions with survivors,
from making them feel uncomfortable, and instead, engaging with them on a human level.
New Journey of the Victim / Survivor and the Police Officer / First Responder
Reflection
The Competition
This project was 1 of 3 projects selected to compete at Interaction 21. Being an undergraduate student from non-design background, I was fortunate to be paired with some of the best interaction design students who gave me a lot of insights and design practices that I never would have come across from a student’s perspective.

3 months of learning working on a very ambitious project, I learned so much about design from the incredible chairs that ran the event, our industry mentors, and the other six design students. Here is a small list of my major takeaways form the competition.
interview, Interview, Interview
One of the best things in this project was the huge number of interview we did continuously. It was prominent how every person’s insights and feedback contributed to the improvement of our ideas and better direction of our solution.

While I was not on the front line of interviews in this project, the flair professionalism of other finalists in my team in this aspect was one of the biggest takeaways in this project.
Keep it human
Being a sensitive topic, it was crucial for us to approach people with whom we as students could communicate and gain the most. A decision on top of this was to not talk directly to victims or survivors, rather advocates who stay close to them and law enforcement officers.

There are hesitations from all stakeholders, including police officers and advocates, not just survivors and victims because of the use of data and processing with it, hence discussing these concerns with experienced officers was useful in positioning and framing of our solution.