July 22 -24
One additional week for additional deliverables presented here

- Researcher
- Designer
- Developer liasion
- Presenter
Skills Employed:

- Design Thinking
- Product Thinking
- Guerrilla Research
- Personas
- Flowcharting
- Wire-framing
- Prototyping

- Figma
- Miro
- Adobe Aero

The Hackathon is a partnership with Walmart Technology and American Pets Alive (AmPA). AmPA is the national education arm of Austin Pets Alive, the largest no-kill shelter nationwide, which aims to save as many animals as possible from unnecessary deaths. This challenge aims to solve problems within the common “Return to Owner” process, when a lost animal is brought to a shelter.


1st Place Winner featured on LinkedIn



WHAT is RTO and WHO are involved?

To design for the RTO process, we first have to understand what is RTO and its process, and the stakeholders involved in this process. Our research relied heavily on literature review and open source information due to the time limitation, but we were still able to construct a working picture of the RTO ecosystem.

How can we help?

After we have familiarized ourselves with the RTO subject matter, we started to research in general current pain points that exist in the RTO process to paint a rough picture of the problem space. In addition, we conducted guerilla user research during the time frame to supplement our findings.

Through our research, we have discovered two most impactful issues , corresponding to the two key stakeholders, within the problem space:

Define + Ideate

Challenge Analysis + Definition

Upon closer examination and further research, we discovered that our previously discovered issues could be viewed as circularly causal: they are both parts of a negative feedback loop of all problems:

And we cannot address these two major issues without breaking the loop entirely, our challenge thereby was defined as coming up with a design solution that can break the any one link of the loop.

Weakest Link + Primary User Analysis

At this point, we decided that a digital solution that could potentially break the loop would need to involve the resources from other stakeholders, as the loop itself stems from the lack of resources of pet owners and shelters. Going back to our intial research, we discovered that Community members/General public has the most extensive conncetions in the RTO process, yet, when circumstances allow, has the most direct connection between the lost pet and the owner.

So by having community members more effectively involving in the RTO process, we can help relieve the crowded shelters, thereby potentially breaking the loop.

Primary User Empahty

After we have defined our primary users, we conducted more research in the forms of literature review and guerilla research.

We were able to conduct shorts interviews with 10 people that indicated that they were willing to help a stray pet. And we have constructed a persona and a workflow to summarize our research.

And we found that there are four main concerns:

Design Goals

Therefore, to help community members, our solution needs to support these high-level goals:

Design Approaches

AI-powered recognition for getting in touch with the owner
- Database for owners to post their pet
- Contact owner in in-app
Easy to follow training and guidance in Augmented Reality for handling animals

These to approaches combined would produce a much more streamlined process, and higher success rate for the community members to return the pet directly to the owner.


Information Architecture

After we have decided on our approaches, we started to determine the informational elements in our design and crystallized that with a information architecture.


After which we started to brainstorm and materialize the design with sketches.

Final Prototype

Home Screen to Accommodate All Roles
In our experience, a user can be just a community member that wants to help, or an owner that wants to post their pets, or both. And our home screen provide relevant information based on their roles.
RTO Training in Augmented Reality
In order to deliver both easy-to-follow and detailed, usable training to our users, we decided to utitlize Augmented Reality. This allows the users to engage with the traning using real physical interactions.

These interactions closely simulate the techniques that would be used in a real RTO scenario, and thereby increasing the intuitiveness and efficacy of the training.

And since the interaction is largely physical, it is more vastly
intuitive and heuristic, meaning the users can readliy focus
themselves in the training itself, rather than wrestling with the interface
There are multiple components in the traning and they react in real time to user inputs
Pet Posting
To allow community members to identify lost pets using their mobile devices, the pet owners would need to post information of their pets onto our solution.

Which includes user-facing informaiton like pictures, description, and system-facing information for computer vision analysis -- video of their pets.
So our experience also guides the pet owners to record videos of their pet in a way that maximizes the result of AI recgonition.
Step-by-step Guidance
The guidance in AR aids the user in helping an animal by giving them the tools they need in heuristic, natural way.

It first walks the users through a guide rail process to determine what kind of animal they are trying to help and ascertain that they are in a position to do so.
As a fail-safe, there is also a chat bot to help with any undefined situations.
Then, it guides the user to try to approach and identify if the animal they are trying help is a known pet right on their devices with relevant AR interactions, without the need for any external tools. This is achieved by leveraging the database of posted pets and computer vision technology.

And if the animal is identified, it provides a direct way to connect with the owner through the app.

Lastly, if the user decides to catch the animal, our solution again provides retrieval guidance in AR in a format similar to the training that the users experienced before, but with added real-time advises based on the animal's reaction to better help retrieval process.

What could be changed + Future Efforts

Due to the 48 hour time frame of the projet, we were not able to receive feedback other than those from the judges of the hackathon, nor produce a interactive prototype. And we are eager to iterate our design with data collected from user testing.

More thorough user research
Once again limited by the time frame of this proejct, we were only able to conduct very basic user research with relatively small sample sizes. And we are fully aware that our extracted insights from the research might be fully presentative of our target audience.

Verify feasbility and efficacy of AI-powered components
While we were designing with our perception of the current state of the techonologies involved, as designers, we were not able to directly confirm 100% feasbility and satisfactory efficacy of our design during the design process.

Zhaohui Zhao