Designing Agentic UX for Salesforce Data Cloud

Shaping AI-native interaction patterns across post-ingestion workflows

Context

Salesforce Data 360 brings disparate enterprise data sources together, but for many customers the hardest part starts after ingestion. Once data is connected, teams still need to understand it, prepare it, and decide what to do next, often across complex workflows and multiple tools.

As generative AI capabilities enter this space, the opportunity is big: reduce manual work and lower the learning curve. But the risk is real too as these are high-stakes environments where teams need clarity, control, and confidence before taking action.

My Role & Focus

As a Principal UX Designer on Data 360, I lead UX strategy for post-ingestion agentic experiences, defining how AI-powered agents support people as they prepare and work with data.

This work spans multiple initiatives, including agent-driven transforms, real-time data profiling, and alignment with broader agent patterns across Data 360, Tableau, Agentforce, and Salseforce’s core CRM.

Contributions So Far

This work is ongoing. My contributions focus on shaping direction across three areas:

Clarifying the role of AI in post-ingestion workflows

I’m defining how agents best support data preparation and analysis, and when users should stay more hands-on. My work focuses on using natural language to help people navigate and execute work in Data 360 in ways that are simple by default and advanced by choice.

Establishing flexible agent interaction patterns

I design and advocate for agent interaction patterns that scale with task complexity. This includes lightweight assistance for quick questions, as well as more immersive experiences for deeper work, rather than forcing a single agent model across all use cases.

Creating alignment across overlapping initiatives

I work across Data Cloud, Tableau, and Agentforce to align teams with overlapping ownership, competing priorities, and pressure to move quickly. I help create shared understanding around agent scope and interaction models, often before formal patterns exist, so teams can move forward without fragmenting the experience.