As Principal Product Designer, I designed a new product feature using artificial intelligence to predict and recommend best practices in data workflows.
AI for Customer Value
The AI model is trained on anonymized mappings across multiple customer tenants. A mapping is a data transformation workflow that moves data from a source to a target with user specified transformations in between. I collaborated with machine learning engineers and UX researchers to identify user behaviors and determine the most appropriate interaction model and user experience for both cloud and desktop products.
The machine learning model determines best practices by analyzing matching sequences within transformations, and based on prior usage patterns produces recommendations for the “next set” of transforms to add during development.
I advocated to integrate the AI recommendations so that the experience was seamless within the existing user workflow, and the best example I used to tell this story was predictive text in text apps.
Predictive text is seamlessly integrated with the existing user workflow and not disruptive and jarring.
Sketches and Renderings
Early conceptual sketches and renderings helped identify conceptual models for the inline recommendations.
In addition to creating interactive prototypes, I use animation to help define interaction paradigms and I find it’s an effective way to communicate concepts to stakeholders and developers.
Early storyboards helped identify interaction models for a panel concept.
Our user-centered design approach identified the primary persona as the Data Integration Developer. Our business objectives introduced another target persona identified by the technology research and consulting firm, Gartner, and that’s the Citizen Integrator.
While conducting user research, the UX Researcher and I observed customers using the newly-released features.
Users think the end results are pretty cool.