2018 Vancouver UX Award Winner

Accelerate

An award-winning augmented reality training application

Role
Agile Product Owner
UX/UI Designer
Team
Alyzah Kaharian, Amy Li,
Charity Principe, Rex Shi, So Jeong Bae

Challenge

How can we make collision estimator training more consistent, engaging and scaleable?

Project kickoff with Accenture

Project
Background

As part of the Masters of Digital Media program at The CDM, our team had the tremendous fortune of working with Accenture, an innovative global management consulting firm. Our initial brief asked us to explore how XR (virtual reality, augmented reality, and mixed reality) could be used to facilitate learning.

What's Your Problem?


Our clients didn’t approach us with a burning problem that needed solving. With the objective to "explore XR technology", we were in some sense asked to explore a solution before we had a problem. There’s not necessarily anything wrong with this approach, there’s definite value in research and exploration, but from our perspective as advocates of human-centered design our brief at times felt anathema to the way things are “supposed to be done”. Were we putting the cart before the horse, in exploring the solution before the problem? We definitely didn’t want to make something for sheer novelty without a valid use-case, but also had to respect that we’d been asked to explore somewhat of a solution.

Given that our clients were open in terms of the specific extended reality medium we would pursue, this meant that we weren’t necessarily required to produce VR for example. In fact, our clients were incredibly gracious and supportive in allowing us to employ design thinking to identify potential problems, and how XR tech might address those issues.

Should we make VR just for the sake of making VR?


So how do we go about finding a problem the XR could solve?

Accenture provides consulting services to a broad range of clients across the globe, including 95 of the Global Fortune 100. For this particular project we would be working with an auto insurance provider. This provided an exciting opportunity and challenge in that we had to work closely with stakeholders from two very different organizations. It also gave us the direct context and potential use-case for what would we would be building.

A secondary objective of the project would be to identify any insights that could be applied more broadly to learning solutions for Accenture’s many other clients across various industries.

Our Approach


Our first challenge and opportunity was in deciphering our brief. Our clients weren’t particularly prescriptive, and showed a lot of faith in us and in the design-thinking process, but at the same time when things are too wide open it can be hard to know what direction to go in. As much as constraints can be challenging, they’re absolutely necessary in guiding the design process. After several meetings with our third-party client, we gained an understanding of some of the circumstances surrounding their current training regimes, and identified collision estimator training as our general problem space. By identifying this space we could now dig deeper to get a better understanding of how estimators were being trained and to surface any pain points that XR technology might be able to address.

Coordinating a Google Ventures Design Sprint with Accenture and our third-party client

Corporate
Training
Problem
Space

  • High cost of classroom training
  • Access to expensive equipment
  • Accesiblity for remote workers/offices
  • Non-standardized training
  • Impact on productivity (on-the-job training)
  • Passive delivery (Powerpoint coursework)

Rapid Prototyping


Over the course of this project some underlying themes started to emerge. One recurring theme is the tension between design-thinking, methodologies or frameworks in a best case scenario, versus design-thinking in the real world. GV design sprints can be incredibly transformative, and by all means, spending 5 days to arrive at a solution is certainly preferable to spinning your wheels for years potentially, but even so, rarely do organizations have the time or resources to allocate an entire week to some problem.

Our clients were extremely generous with the time and energy they committed to the project, but as could be expected, setting aside an entire week for a student project would not be feasible. So we had to improvise. We managed to condense the Google Ventures Design Sprint into a single day, and while this is far from optimal it nevertheless yielded some valuable insights. If anything this is a testament to the strategy, that even in such a truncated capacity, it can nevertheless be revealing.

As an outcome of our micro-sprint, we identified X number potential divergent prototypes that warranted further exploration and interrogation.

Paper prototypes for review and critique

3 Prototypes

AR in the classrom

Use augmented or mixed reality to enhance existing learning materials or as a focal point for instruction.

VR Papers Please

VR simulation of the estimation auditing process that would train for pattern-matching and critical thinking.

AR Estimate Builder

Use augmented reality to train collision estimation on a virtual vehicle

How Will We Decide?

With a problem space defined, and multiple solutions prototyped, our next challenge was defining scope, and identifying which direction we would pursue. As this was only a 13-week project cycle, we had to be realistic in terms of scope and fidelity. We could go broad and wholistic, and deliver something conceptual or proof-of-concept, we could niche down and build something real and high-fidelity, or we could deliver some combination thereof. While our own personal biases pulled us toward building something “tiny and robust”, we had to defer to the needs of our multiple client stakeholders. We had our own personal preferences in terms of intrinsic motivation and interests that could interfere with decision making, and understandably clients too might be drawn to some potential direction for subjective reasons. We needed to define some objective decision making criteria with which to identify which direction we should pursue. In our brief and in previous discussions with our clients, we identified several ideal requirements, but given the short time-frame, how many of these were nice-to-haves, and how many were critical. We identified any requirements that had been mentioned thus far and using a card-sorting exercise asked our client to identify which four of the potential requirements were most important. Of these, our client identified to following as critical project criteria:

Design Criteria

Good AR Use Case

We were not interested in developing something purely for novelties sake. It was we imperative that we identify a solid use-case for the XR technology. We frequently asked our selves, “is this something that could be better solved by a poster, video, website?” Does XR actually solve a problem.

High Deliverable Fidelity

Accenture is a respected global organization, and as much as they understand the design process and can look past the fidelity of a prototype or MVP, the intention was to demo this project to clients, so it had to have professional level of fidelity.

User Research Validated

It was important for us to work with who would be the end-users of this product

Feasible to Implement

While this project was exploratory in capacity, we did not want to create solution that would be impossible to implement

With these design criteria clearly defined and prioritized, we then ranked our three prototypes in terms of how they met these criteria. Of the 3 potential directions, the estimation simulation was identified as the optimal direction to pursue

Researching and documenting the estimation process

User Research Deep Dive

With estimating identified as our problem space we now had the opportunity to dive deep to understand the space and our users. We observed users perform several estimates. We had interviews with several estimators as well as users up and downstream of the estimate process. Through the course of this research we learned that photography is a critical business function. In some cases estimators work entirely off of photographs provided by third-party collision repair service providers. For estimates performed in-house these photos are critical documentation in terms of review and auditing, and can be submitted as evidence in court-proceedings. Additionally, there are compliance requirements for the order that photographs are entered into the software system. While there are standards for compliant collision photography, there are nevertheless poor photographs and the current training for photography consisted of a one-page PDF. During research interviews with estimator trainees, we discovered that many of these trainees self-identified as kinaesthetic learners, meaning that they learned best from hands-on training. Hands-on training can be costly and challenging to implement, it requires an inventory of damaged vehicles, as well as trainers to administer and educate. In a large geographical area, it requires travel to a central facility that may be hundreds or thousands of kilometres away.

Estimator training pain points

Friction Point How Might We Solution
Geography Train remote estimators Mobile application
Passive Learning Better engage student Immersive gamified experience
Access to damaged vehicles Provide better access High-fidelity, life-size models
Inconsistency in training Standardize learning Self-guided standard lessons
Hands-on training Virtual hands-on Interactive augmented-reality
Ongoing training impacts productivity Train remote estimators Self-guided supplementary training

Our Solution

An immersive virtual hands-on training simulation for the next generation of estimators

Trace the steps of an estimator

Follow guidance to get into the right position to capture your photo

Immersive

Experience the immersion of Virtual Reality with the accessibility and portability of AR

Active Learning

Perform the actual movements of a real estimator

Location Tracking

Guidance through the virtual world

Course Platform

The augmented-reality portion of the application is situated within a larger self-guided training platform, photo training tool would be one of many course modules

Course platform flow prior to launching augmented-reality lesson. Using Blooms Taxonomy for classifying learning objectives.

Post engagement summative feedback and gamification.