The Trade Desk is a demand-side platform (DSP) that enables advertisers to buy and manage digital ads across display, video, audio, mobile, native, and connected TV. With the launch of Kokai, The Trade Desk introduced AI-driven tools to help advertisers make smarter audience decisions. Instead of relying on manual searches and broad personas, which often lead to wasted ad spend and missed opportunities, Kokai provides real-time insights that predict which targeting strategies will deliver the best results.

Advertiser Seeds helps predict targeting success before a campaign launches. By analyzing past converters from first-party, third-party, or contextual data, seeds help generate relevance scores that grade audience quality. These scores help advertisers identify the best-performing data sources, refine targeting, and optimize ad spend.

My challenge as the Lead UX Designer was to design an intuitive, transparent experience for creating, managing, and activating Advertiser Seeds.

Key contributions

• Defined the core user experience for Seeds within Kokai
• Designed end-to-end workflows for Seed creation, management, and activation
• Collaborated with Data Science to translate AI insights into clear, actionable relevance metrics
• Conducted user research to validate and refine the experience
• Aligning with Product, Engineering, and Business teams to ensure seamless integration

Please email for full case study.

Finding the right audience isn’t as easy as it sounds. Advertisers have access to hundreds of thousands of third-party data segments to build audiences with, but:

• There’s no clear way to evaluate data quality—they often rely on intuition.
• Segments don’t always match expectations —a category like “Luxury Shoppers” might not actually contain their target audience as it is too generalized.
• Campaign performance suffers because advertisers struggle to find the best audience fit.

With Kokai shifting from goal-based buying to audience-based buying, we needed to make Seeds the foundation for smarter, more precise audience selection.

Our hypothesis is that with Seeds, advertisers could stop guessing and start making data-driven decisions.

By leveraging AI, Seeds provided:

• A relevance score to evaluate the quality of each audience segment
• Data-driven insights to compare different targeting options
• A seamless way to activate the best audiences

This shift had the potential to dramatically improve targeting accuracy and campaign efficiency—but only if we could make the experience intuitive, insightful, and easy to adopt.

Mapping user touch points and workflows

To create an intuitive experience, I first needed to understand:

• Where in the workflow advertisers would engage with Seeds
• What information they needed to trust AI-driven recommendations
• How they would evaluate and activate audiences

I worked closely with Product and Data Science teams to document every Seeds touchpoint—from data input to AI-driven recommendations.

Structuring the Seeds Experience

Through the seeds mapping exercise, I identified three key steps in the advertiser workflow:

1. Creating Seeds – Advertisers define their converters (users who have interacted with their brand).
2. Evaluating Data – AI engine assigns relevance scores, ranking segments based on quality and value.
3. Activating Audiences – Advertisers select high-quality segments to improve targeting.

I designed the UI to guide users seamlessly through each stage, ensuring:

• A frictionless setup process for creating Seeds
• Clear, digestible AI insights to simplify decision-making
• A natural integration into audience selection workflows

Designing for Transparency & Trust

One of the biggest challenges was ensuring advertisers trusted the AI-driven recommendations.

To address this, I focused on clarity and control:

• Explained how relevance scores were calculated—instead of presenting AI as a black box allowed users to adjust parameters—giving them flexibility in audience selection
• Provided direct comparisons—showing why one segment was stronger than another

This ensured advertisers felt empowered, not replaced, by AI-driven insights.

Testing & Iteration

I conducted multiple rounds of usability testing, where users:

• Created Seeds and evaluated audience quality
• Interacted with AI recommendations
• Provided feedback on clarity, usability, and trust

Key insights led to:

• More intuitive relevance scoring—breaking down metrics into digestible insights
• Guided recommendations—highlighting best-fit audiences first to simplify selection
• Progressive disclosure—revealing details only when needed, preventing information overload

The MVP launch of Seeds introduced:

• A seamless Seed creation and management experience
• AI-powered relevance scores for data quality evaluation
• A guided process for audience creation and selection

Advertisers could now make smarter, more confident decisions when selecting audience data—removing guesswork and improving campaign performance.