Our overall goal is to use a machine learning system to deliver the right message at the right time to a user in order to improve their activity level. A conversational agent will be used to handle the communication to the user, which will be delivered via a mobile application.

is to help people become more active by combining:

• A Conversation Agent
  ‍The conversation agent is the core interface between the user and the underlying data that feeds the machine learning models.
• Machine Learning Nudge Engine
  ‍The nudge engine is the mechanism that ingests data, learns, then feeds the relevant prompts back to the user.
• A Mobile Application
  The mobile application is the landing place for the information and is dispersed into the relevant areas for end user consumption and for the conversation agent to interact with the user.
• Data Platform
  ‍The system is backed by a platform that manages the inputs and provides research access to the data.

Out team consisted of a project manager, a full-stack app developer, a UX/UI Designer (me) and two engineers who specialised in machine learning platforms, nudge engines and engagement models. In order to achieve our goal, the project was to divided into three phases: Foundation, Iteration and Extension.
‍
Most of my efforts were executed in the foundation phase. While the developers were setting up the backend architecture and supporting environments, it was important that I charged in with executing my design discovery as efficiently as possible so the developers could start creating the app and inputting the conversation flows.

The iteration phase was used to expand, diversify and test the conversation discovery, while iterating the app's wireframes according to user feedback.

• Phase 1: Foundations
  - Foundation of Nudge Engine, data mapping and ML Pipeline.
  - UX design process to create the engagement model, agent personas, dialog and mobile application.
  - Backend architecture and data pipeline.
  - A working iOS app to collect data and facilitate both the nudge engine and conversation agent.
• Phase 2: Iterations
  - Involve users Involve users in a test cycle to test nudge engine and engagement models.
  - Develop Agent’s ability to engage users and its depth of situational awareness.
  - Develop the App to accommodate more data and increase user engagement
  - Iterate on on Nudge Engine, exploring different models and data requirements
• Phase 3: Delivery
  - Continue to tweak the Nudge Engine, improving accuracy and work on data model efficiency.
  - Potentially add additional activity monitors and platforms to extend reach.
  - Create a Research dashboard for monitoring and supporting the platform.
  - Prepare the platform for the research phase and the ability to run self sufficiently for 2 years

Design Discovery

• Uncovering user archetypes, user flow, empathising. Uncover insights into the highest value functions with the greatest impact. 

• Validate initial assumptions in a collaborative design workshop with end-users and key stakeholders. 

Wireframing & Prototyping

• Design of application(wireframes) that is in line with UI best practises, and commonly used web platforms.

• Apply to a clickable Prototype that shows key interactions.