Omni Assist
Using AI to simplify complex workflows and move users from intent to action.
Scope
Director, UX/UI
Role
UX Strategy, AI Experience Strategy, Product Design, Design Leadership
Platform
Web / Generative AI
The Problem
Omni gave marketing teams access to powerful data, insights, and tools, but finding the right information and knowing what to do with it often required navigating across multiple products.
As the platform became more complex, users needed a faster way to find insights, understand what they meant, and take action.
Generative AI created an opportunity, but simply adding a chatbot wouldn't solve the underlying problem.
The challenge was to determine where AI could meaningfully reduce complexity within existing workflows.
Users needed:
Faster access to relevant insights
Help interpreting complex information
Clearer paths from insight to action
Guidance without losing control of important decisions
The Opportunity
Rather than treating AI as another feature or destination within Omni, I saw an opportunity to use it as an intelligent experience layer across the platform.
How might we allow users to begin with what they're trying to accomplish instead of requiring them to understand where every capability lives?
Our design goals were to:
Make complex platform capabilities easier to access
Connect insights directly to recommended actions
Guide users through multi-step workflows
Establish AI interaction patterns that could scale across Omni
Research
Before defining the experience, I worked with Product, Engineering, Data Science, and Design to identify where AI could provide meaningful value within existing workflows.
We examined:
Common questions users repeatedly needed to answer
Workflows requiring multiple Omni products
Repetitive tasks that could be simplified or automated
Moments where users had information but didn't know what to do next
The context users needed to understand and trust AI responses
This helped us focus the experience around real workflow problems rather than introducing AI simply because the technology was available.
Insights
The research revealed that users weren't thinking about which Omni capability they needed.
They were thinking about their goals:
Why is my campaign underperforming?
Which audience should I target?
What needs my attention?
This shifted our thinking from AI as a conversational interface to AI as a way to connect user intent with capabilities across Omni.
That became the foundation for Omni Assist.
Key Insights
Users think in goals, not products
Users often needed to know which application contained the information or capability they needed before they could begin.
Opportunity: Allow users to start with their intent and let Omni determine which capabilities can help.
Insights alone weren't enough
Providing an answer still left users responsible for interpreting the information and determining what to do next.
Opportunity: Connect insight → recommendation → action within the same experience.
A blank chat box wasn't inherently simple
Open-ended conversation gave users flexibility but assumed they already knew what Omni Assist could do.
Opportunity: Combine conversation with suggested prompts, recommendations, and guided workflows.
Explorations
The first direction focused on helping users find information across Omni through natural language.
It made information easier to access, but the experience still stopped at the answer.
What we learned: Finding information faster solved only part of the problem.
Intelligent Search
We explored a more traditional conversational model that allowed users to freely interact with Omni Assist.
While flexible, users didn't always know what they could ask, and complex workflows were difficult to communicate entirely through chat.
What we learned: Flexibility needed to be balanced with guidance.
Open Conversation
The final direction combined conversation with structured product experiences.
Suggested prompts helped users begin, recommendations provided relevant next steps, and traditional UI appeared when greater precision or control was needed.
What we learned: The strongest experience wasn't chat alone. It was AI guiding users from intent to action.
Guided AI Workflows
The Solution
Omni Assist became an intelligent experience layer that connected users to insights, recommendations, and actions across the platform.
Key Features:
Contextual prompts & Natural-language interaction
Suggested prompts showed users what was possible, while allowing them to start with a question or goal instead of navigating multiple products.
Insights + recommendations
Responses helped users understand what was happening while recommendations provided clear next steps.
Guided AI workflows
Omni Assist could move users from a question through a structured workflow, bringing in traditional UI when greater precision or control was needed.
Progressive disclosure
Complex information was surfaced gradually, allowing users to explore additional detail without overwhelming the initial experience.
Designing an AI System
Omni Assist needed to scale beyond a single use case.
I led the definition of reusable AI interaction patterns for prompts, responses, recommendations, guided actions, system states, and transitions between conversation and traditional UI.
I worked closely with Product, Engineering, and Data Science to balance user needs with the evolving capabilities and limitations of the technology.
This gave teams a shared foundation for introducing new AI experiences while maintaining consistency across Omni.
Outcomes
Users could begin with a question rather than navigating across multiple applications.
Faster paths to insight
Contextual prompts and recommendations surfaced tools and insights users may not have otherwise discovered.
Better capability discovery
Insights, recommendations, and actions became part of one continuous experience.
More connected workflows
Reusable interaction patterns established a foundation for expanding AI and agent-based workflows across Omni.