WDA: Government Application

One of the biggest challenges early on was access. Due to the sensitivity of the work and the security clearances required, we had limited direct contact with our target users. To overcome this, we proposed an off-site design thinking workshop as a low-risk way to begin collaboration. The workshop was a turning point—not only did it provide valuable insights, it also helped us earn the trust of stakeholders.

Following that workshop, we were invited on-site to conduct in-depth user interviews and field research. These sessions were conducted under tight security conditions—no phones, laptops, or cameras allowed—forcing us to be creative in our methods. Over the course of our visit, we interviewed about ten users across various roles, observed their workflows, and ran contextual exercises to understand how they collect, evaluate, and synthesize information.

The research culminated in a set of detailed journey maps that we presented to users, engineers, and stakeholders. These maps were instrumental in aligning the team and shaping the scope of Release One. We used the insights to define key user needs, prioritize features, and structure our sprint cycles.

WDA: Pharmacology Application

In parallel, we developed designs tailored to pharmacological research, focusing on uncovering molecular and genetic relationships. One wireframe prototype explored how a researcher studying TP53—a gene that produces a tumor-suppressing protein—might search, disambiguate, and trace connections through WDA.

Our interface included a smart query bar that suggested relevant terms and prompted disambiguation before executing a search. This design decision was backed by research across both verticals: researchers frequently encounter acronyms or terms with multiple meanings, and clarity up front increases trust in the results.

Search results landed users on a dynamic overview page, surfacing the target entity and its related concepts. Users could then explore connections, view supporting evidence, and investigate further through a rich data visualization interface. They could also apply filters, save graphs to research projects, and "watch" a question—receiving alerts when new research matched their area of interest.

Cross-Product Collaboration

As a small design team embedded in a highly technical environment, we worked closely with dozens of engineers developing the Watson knowledge graph, as well as with product managers, marketers, and legal teams. The core challenge was the complexity of the underlying AI technology—how Watson ingests, organizes, and surfaces information from a custom-trained corpus. To design effectively, we had to deeply understand how the knowledge graph operated, without letting that complexity surface in the product experience.

Our role was to translate the sophistication of the system into a clear, intuitive interface. This meant ongoing collaboration with engineers to understand system constraints and possibilities, and close alignment with product teams to ensure we were solving the right problems for researchers. The success of the product hinged on our ability to maintain this balance: respecting the depth of the technology while delivering a streamlined, approachable experience for the end user.

Success & Impact

Watson Discovery Advisor (WDA) was one of the first commercially available applications to leverage IBM Watson’s AI capabilities for real-world problem solving. As a product, it demonstrated the power of cognitive computing in specialized research domains, particularly in pharmacology and government intelligence. WDA allowed researchers to uncover hidden relationships between entities such as genes, proteins, individuals, and organizations—surfacing insights that were previously buried in massive, unstructured datasets. The application helped validate the commercial potential of Watson’s knowledge graph and natural language understanding technologies, paving the way for future enterprise use cases and influencing the design of the Watson Discovery API, which today powers multiple IBM products. WDA's success also helped establish credibility for human-centered design practices within IBM, showing how thoughtful design and AI could work hand-in-hand to deliver meaningful results.