Aiva - Amazon IT Services

AIVA Amazon IT Services

Aiva, Amazon's AI-powered IT assistant, uses machine learning to identify and troubleshoot IT issues, offering self-service solutions or directing employees to live support, reducing direct IT intervention.

I was tasked with figuring out how to optimize the Aiva experience, coming up with numerous deliverables and high-fidelity clickable prototypes within a 4 month timeframe.

To comply with my non-disclosure agreement, I have omitted and obfuscated confidential information in this case study. All information in this case study is my own and does not necessarily reflect the views of Amazon.

Background

The Amazon IT Service internal portal was designed for corporate Amazonians to address their device or software-related issues. Aiva was a chatbot connected to the IT portal that, when clicked, attempted to help solve Amazonians IT issues.

The problem was Aiva was not discoverable on the internal page, leading to usability challenges, and she only served as a reactive support method.

Current IT Service Portal. Aiva appears upon clicking "Get Support" button

THE CHALLENGE

Enhance AIVA IT Support Experience

My goal was to revamp Aiva to enhance overall IT Support Satisfaction. The original premise of Aiva was simple: make it as easy and quick to solve your IT problem with the least amount of resources/clicks.

My high level goals were to:

Enhance Aiva experience/IT service experience
Reduce the number of clicks users had to reach their solution
Create a platform for higher quality user engagement and build trust with IT brand

MY ROLE

I led the design experience and research between June 2022 - March 2023 and collaborated with five other designers and researchers on the IT Service team.

I also worked alongside 2 Product Managers and reported to a senior designer and design manager. I pitched my designs to the Senior VP of the IT Services division and other Director stakeholders. My designs ended up informing a company wide 3-year plan and got pushed into production.

Understanding the Journey

Initially, I didn’t have a clear mission or specific goals for how exactly I wanted to approach enhancing Aiva. Without these insights, I delved into some research to explore how users were utilizing the IT Service backend portal.

EARLY INSIGHTS

To understand the IT Service pipeline, I engaged in some preliminary testing with my own computer and attempted to troubleshoot some on-boarding issues. Based on various test cases I navigated through, I constructed a rough pipeline of what the average Amazon employee goes through when they engage with IT.

1/1

THE DISCOVERY

Amazonians sought IT Support for 2 Types of Issues

Of the users that utilized Aiva/Amazon IT Service, I could break them into two groups:

High productivity impact issues - a problem requiring immediate assistance affecting an Amazonian’s daily productivity
Low productivity impact issues - an issue which could be resolved later but not urgent to the productivity of the user

This revealed an opportunity to craft a more positive IT experience for everyone, no matter what role they were in–a working north star.

DEEPER INSIGHTS

Research and More

Before I could start designing, I needed to understand what other competitors were doing to enhance their service pipelines with chatbots. I looked at Apple, Kayak, T-mobile, and Bank of America websites to understand their pipeline to resolving issues and utilizing chatbots as a medium to troubleshoot customer problems.

I discovered that most utilized chatbots as a way to minimize support calls from live service representatives. These bots served as personalized FAQs but were inefficient when it came to unique or rare issues, and people gravitated toward contacting the live service rep anyway. Apple’s pipeline didn’t use a chatbot. They had the user filter down their issue and provided helpful articles. If they still needed help, they were directed to a live service rep.

1/1

Apple IT Service Pipeline - no chatbot, users clicked through categories that prompted helpful articles for troubleshooting. If no solution was found through self-service, they were directed to a live service representative.

After understanding this, I decided it would be important to define success and understand the Amazon IT Service experience from beginning to end.

To do this, I had to understand the concept of the perfect IT service pipeline and what mindset our users were in starting from before they encountered a problem to after, accounting for stress, satisfaction, and time spent to completion of the problem.

Based on these new user flows, I was able to envision Aiva as a tool that could be used as both proactively AND reactively to better educate users, identify problems, and solve user issues in the overall Amazonian daily experience.

Users' Initial Mindsets Were Negative from the Outset

Delving deeper revealed some incredible insights into the overall IT Service experience. When engaging with IT, users already had a negative bias and entered the process with a frustrated mindset. In fact, previous research on Amazon IT support revealed that customers perceive IT as “punitive, policed, and matter of fact.” It also didn’t matter how tech-savvy they were or not; Most also didn’t know Aiva existed, let alone what the tool was called.

Most importantly though, users used Aiva as a reactive support method, when Aiva’s ML capabilities should be leveraged to provide preventative support. The time and energy users had to go through to resolve a small IT problem (70% of IT Problems) was not worth the degree of the problem. So, I went about identifying a solution.

REFRAMING THE PROBLEM

Inefficient Touchpoints Lead to Downstream IT Problems and Reinforce a Negative Brand Image.

Through my research, I discovered that the inefficient placement of Aiva along the entire IT service pipeline worsened the experience for all users while increasing demand on IT service reps / reinforcing negative associations with Amazon IT brand.

So, how might we better place Aiva along the IT service pipeline to optimize her capabilities and improve IT satisfaction?

User Journey Sketches. Attempting to better understand overall IT Service experience

CUSTOMER JOURNEY MAP

My proposal was Aiva as a preventative and proactive IT service tool. To see where I could improve efficiencies in the service pipeline, I created several customer journey maps with high productivity and low productivity issues to reveal gaps in self-service and brainstormed potential Aiva touchpoints

Collaborating with 2 PMs, we broke down the 10 most common IT issues and labeled them high/low productivity impact issues based on the amount of time it typically took to resolve the problem.

Potential touchpoints for High Productivity Impact Issue - Password expiration.

For the high productivity impact issue, we settled on password expiration, a use case in which if the Amazonian’s password expires, they get locked out of their device

For the low productivity impact issue, we decided on hardware replacement request, a use case in which if an Amazonian is eligible to upgrade their laptop/device, they request it through Amazon IT.

I envisioned Aiva as a tool that could be used both proactively and reactively to educate users, identify problems, and solve user issues in overall Amazonian daily experience with these potential touchpoints:

1. Tamagotchi Aiva - IT health indicator, gamification for user engagement, brand positioning

2. 1:1 meeting pop up - Educational, proactive approach

3. Virtual pocket assistant - Better reactive support, in case of emergency

THE REDESIGN

Introducing the IT Desktop App

An all-encompassing IT service desktop app that acts as a reactive solution tool as well as a proactive educative tool with personality and gamification to enhance overall user experience.

I pinpointed these three design goals that I wanted my wireframes and prototypes to address:

1. Increase Aiva's adoption as the primary IT support resource by fostering trust.

2. Position Aiva at key touchpoints to better prevent and diagnose IT issues.

3. Refine Aiva’s personality, appearance, and placement to improve perceptions of Amazon IT Service.

INCLUSIVITY & ACCESSIBILITY

Design Workshop

To quickly ideate on Aiva's design and counter users' negative perceptions of IT, I hosted a Design Jam with 5 other designers and researchers. Our goal was to create a more humanized and approachable bot and brainstormed different design concepts keeping in mind cultural and continental differences.

We started by defining Aiva's purpose as helping Amazonians solve their IT issues as quickly as possible and mapped out her range of emotions that she would experience if she were human.

1/1

We then chose the top main adjectives that we thought most embodied her personality as an IT Support assistant and rough sketched some ideas on what we thought she should look like based on these words.

1/1

The most popular concept was Aiva as a dog/pet, which resonated well due to Amazon's pet-friendly culture and offered a positive, friendly way for users to engage with their IT issues, similar to taking care of a pet. So we modeled Aiva as a pet with a range of emotions that would gamify the user-experience.

1/1

A MORE INCLUSIVE DESIGN

Giving Employees Back their Time

I translated the conceptualizations of Aiva to wireframes that addressed two use cases (High Productivity and Low Productivity Impact Issues) and focused on three features/concepts of Aiva:

1. Aiva as a gamified experience
2. Aiva as an educational tool
3. Aiva as a proactive IT agent

1/1

After a round of critique meetings from senior designers and managers, I iterated on the design concepts using Amazon Design systems to create a more simple, eye-catching and easy-to-use interface that more accurately displayed Aiva's personality and brand image.

1/1

FINAL DESIGNS

Clickable Prototype

For the final designs, I used my previous user flows of the high productivity impact issue (password expiration) and low productivity impact issue (hardware replacement request) and added an educational “Tip of the Day” user flow to showcase the functionality and concept of my IT desktop app.

Presentation Pitch

Impact

Pitched to Senior Leaders and Stakeholders, earning a return offer to continue with usability testing and user research.

Influenced a 3-year company-wide plan which greenlit app into development.

What's Next?

After receiving the return offer, I planned the next four months of my continual internship to:

1. Conduct more user research to define user conceptions of productivity

2. User test to further validate Aiva’s touchpoints along the user journey and determine the most efficient features that solve both high/low productivity impact issues

3. Explore Aiva as an all encompassing support assistant in-app experience

4. Investigate impact of variations of Aiva’s personality

5. Explore cultural perceptions of Aiva from users of different cultures in order to promote inclusivity and diversity