We have a problem

Inaccurate data, approval delays, and fragmented workflows made timely public-health marketing difficult.

27+ interviews were conducted asking questions on AI use, productivity, workflows and features. Above listed were the top insights from the user interviews.

So,do they use AI?... Answer is Yes.

Feature Priortization: Chat, Text based generation, Sources, Text Editor

16 users marked most common tasks based on importance, difficulty, control, time spent, AI openness.
This guided us in the Feature prioritization of the MVP. Over the course of this project, we talked to approximately 40+ public health officials.

Results from interviews and task ranking activity confirmed our intuition. Writing and source-checking ranked the most important features with underlying themes of AI bringing efficiency, transparency and user control in the workflow.

Solution space

Not re-inventing the wheel

We adapted known editing and content-review patterns to reduce learning curves while keeping AI interactions transparent.

Solution space

Trust has legs

“Trust” is not a vague aspiration. It is moving, evolving and built with small moments through out the journey. It's a business promise, tied to KPIs, supported by features, and validated through instrumentation.
UX decisions were guided by design principles and user feedback.
Product decisions created a pipeline of execution to fulfill the KPIs of trust.

Validation

Feedback and improvement

We moved quick. We built a Figma prototype with mid-fidelity wireframes to capture insights with an enthusiastic pool of users. We received general usability feedback but showing the AI feature specific feedback that were added to the subsequent iteration.