IBM WATSON - GOING BEYOND A.I.

IBM Watson is taking Artificial Intelligence to the next level by building cognitive systems that can Understand, Reason, Learn and Interact with users. While most people may think that Watson is a computer, it is actually an open set of cognitive APIs which can be combined to harness various areas of expertise. Watson learns by being trained in various areas and allows for customization which often times sets it apart from competitors.

Our team worked out of the IBM Design office in Austin, Texas.

Learn more about IBM Watson

My Role: Team Lead/Designer/Researcher (Watson Speech and Language Services) | Jul '14 - Jan '16
Skills Gained: Practical User Research and Persona Development, Translating skills from Industrial Design to User Experience Design, and working in a 400k+ person organization.

CUSTOM TRANSLATIONS FOR A CUSTOM WORLD

The IBM Watson Language Translation Tool allowed users to train Watson to produce more accurate, domain-specific (initially law, medicine or finance) translations utilizing Machine Translation without the need for Human Translators. The goal for this tool was to allow users (i.e. Project Managers) to upload data to train Watson without extensive knowledge of a language or technology, reducing turnaround time and expense.

This web-app was the first Watson customization tool that allowed non-developer users to interact with an API. Throughout the project, we worked with Watson API Researchers and Sponsored Users from two Translation Firms based in the U.S. and Europe to develop a tool that worked faster and smarter for our targeted user group. We traveled globally to meet with potential users and to make changes quickly which allowed us to move swiftly from idea to Open Beta in November of 2015. Unfortunately, this product was shelved before General Release in 2016 after I left IBM and the team dissolved. 

Release Blog

Concept to Release: 7 months - Open Beta
Team Size: 2-3 Designers, 2 Developers, & Many Researchers
Designers: Josh Tahmasebi (Team Lead), Britany Ponvelle (Visual Designer), & Jo Yee Pang (Research)

HELPING MAN AND MACHINE UNDERSTAND EACH OTHER

The Language Translator (previously Language Translation), Speech to Text and Text to Speech services are three projects that I managed and helped to design (the service and educational assets) as they moved from Beta to General Availability.

Our team evaluated these services from end-to-end to ensure that they not only provided added benefits to the user, but were simple enough for users to understand how they could use that service. To accomplish this, we designed demos and landing pages as educational aides for each service. We also helped to standardize a series of educational videos about the services (still in use at IBM today) which not only explained the service, but gave potential example of use cases. All of this was accomplished while being mindful of potential target markets and designing tools to customize each service.

Learn more: Language Translator  |  Speech to Text  | Text to Speech

Try Language Translator  |  Speech to Text  | Text to Speech

Concept to Release: 2 months
Team Size: 2-3 Designers, 1 Developer, & Many Technical Researchers
Designers: Josh Tahmasebi (Team Lead), Britany Ponvelle (Visual), & Jo Yee Pang (Research)

OUR SHOWROOM OF COGNITIVE SERVICES

The Watson Developer Cloud was launched in October 2014. The website site is the main hub for learning about various Watson services, as well as, trying each service. Shortly after the site launched, I joined this project.

Upon joining, my initial role was to determine users were being driven to the information they were looking for. Coming at this with fresh eyes, as well as, working with users and performing a heuristic analysis, allowed us to realize that there were some problems. In my findings, I discovered that users were not sure how to get started with using the services, where to use the services, or what the general idea behind cognitive computing was. The website seemed to much like magic, rather than breaking it down in easy to understand chunks.  Within two weeks of that initial launch, we were able put out a revised site rectifying these issues and had a measured effect on our potential users understanding. After that, the vast majority of my work centered around creating demos for new services and creating "Gallery Apps", which are demos using the services in practical applications. We also were trained in and utilized A/B testing tools that could be utilized to further our research.

Watson Developer Cloud
Demo Application: Celebrity Match

Time on Project: 8 months (September '14 - April '15)
Team Size: 2 Designers & 1 Developer
Designers: Josh Tahmasebi (UX), Leslie Shapiro (Visual), & James Zhang (Developer)