THE CHALLENGE

Designing an AI-powered contextual advertising platform for the post-cookie era
With the decline of third-party cookies, advertisers needed new ways to deliver highly targeted campaigns while maintaining brand safety and consumer trust.

Silverbullet launched 4D — a contextual advertising and insights platform that used machine learning, semantic analysis, and computer vision to help brands target audiences through content rather than personal tracking.

I joined shortly after launch at MVP stage as the company’s first UX/Product Designer, working across research, product strategy, UX, design systems, experimentation, and feature delivery alongside Product, Engineering, and Data Science teams.

Over a two-year period, we evolved the platform into a scalable product ecosystem used to:

- Build contextual advertising campaigns
- Surface AI-driven optimisation insights
- Automate performance recommendations
- Support advanced video targeting and partner integrations

THE PRODUCT

The platform combined contextual targeting, optimisation tooling, and machine learning-powered insights into a unified workflow for media planners and advertisers.

Core capabilities included:
- AI-assisted contextual targeting using semantic analysis and computer vision
- Real-time optimisation insights and recommendation systems
- Performance reporting and campaign analytics
- Video placement targeting across premium inventory including YouTube
- Partner integrations through a modular “Dimension Marketplace”.

The product was designed to simplify highly complex advertising workflows while enabling faster campaign setup, improved targeting precision, and scalable optimisation.

THE PRODUCT | SCALING THE PLATFORM

When I joined, the platform consisted of a small number of core workflows focused primarily on contextual campaign setup. Over the following two years, we helped evolve the product into a much broader ecosystem supporting:

- Campaign creation and management
- AI-assisted optimisation
- Reporting and insights
- Notifications and workflow tooling
- Modular targeting systems
- Partner integrations and marketplace capabilities

As the platform expanded, a major challenge was maintaining usability and navigational clarity while introducing increasingly complex functionality across multiple user workflows. To support this growth, I introduced more scalable information architecture patterns, reusable UX systems, and modular navigation structures that enabled the product to scale without significantly increasing cognitive load for users.

VIDEO TARGETING | AI-POWERED CONTEXTUAL ADVERTISING

I helped design advanced video targeting tools powered by computer vision and machine learning models. Unlike traditional contextual advertising platforms that rely heavily on metadata or audio transcription, 4D analysed video content frame-by-frame to better understand topics, sentiment, faces, brands, and contextual relevance across premium inventory, including YouTube.

Working closely with Data Science and Engineering teams, I translated complex targeting capabilities into scalable, user-friendly workflows that allowed advertisers to:

- Target or block topics, brands, and faces
- Build contextual targeting strategies visually
- Manage large-scale targeting configurations efficiently

The result was a more accessible AI-powered targeting experience that improved targeting precision and campaign relevance across video advertising environments.

AI & AUTOMATION | ACTIONABLE AI INSIGHTS

We designed an AI-assisted optimisation system that transformed complex campaign data into actionable recommendations for advertisers and media planners. Working closely with Product, Engineering, and Data Science teams, I helped simplify highly technical machine-learning outputs into intuitive workflows that allowed users to:

- Identify high-performing environments
- Understand campaign performance drivers
- Automate optimisation decisions
- Improve targeting efficiency in real time

The challenge was balancing powerful AI-driven capabilities with usability for non-technical users. To solve this, I designed simplified recommendation flows, prioritised insights, and guided optimisation interactions that made complex data easier to understand and act upon. The feature contributed to significant campaign performance improvements, including 200%+ CTR uplift across optimisation-focused workflows.

STRATEGY 

In the first couple of weeks of joining, we held several online workshop sessions to gain an understanding on the immediate and long term business goals of the company. Processes for each departments were refined.

STRATEGY | PRODUCT VISION & ROADMAP

To align the platform with both immediate business needs and long-term product goals, we ran a series of cross-functional strategy workshops during the first weeks of the project. Stakeholders across Product, Engineering, Data Science, and UX collaborated to:

- Define the product vision
- Identify technical opportunities and constraints
- Prioritise roadmap initiatives
- Align teams around measurable business outcomes

These workshops established a shared direction for the next 12 months and helped guide platform decision-making across teams.

RESEARCH

To better understand user workflows and operational pain points, I conducted interviews and usability testing sessions with account managers across the US, UK, and Europe.

Sessions focused on:
- Campaign setup workflows
- Contextual targeting experiences
- Search and keyword interactions
- Reporting and optimisation tasks

I also reviewed competitor platforms including Peer39, Grapeshot, and GumGum to identify usability gaps and opportunities within the market.

RESEARCH | USER INTERVIEWS & TESTING

The research revealed that many advanced platform capabilities were difficult to discover and overly complex for everyday users. These findings directly informed improvements to navigation, targeting workflows, and campaign management experiences across the platform.

RESEARCH | SYNTHESISING USER FEEDBACK

Research findings were synthesised into recurring usability themes, workflow pain points, and feature opportunities. These insights directly informed prioritisation across navigation, targeting, reporting, and optimisation experiences.

RESEARCH | IDENTIFYING FRICTION ACROSS CORE WORKFLOWS

User interviews revealed recurring friction across campaign setup, contextual targeting, search, and optimisation workflows. While users valued the platform’s advanced capabilities, many features were difficult to discover, overly technical, or inefficient for day-to-day campaign management.

Feedback was organised into positive experiences, usability concerns, and feature opportunities to help prioritise improvements across the product roadmap.

RESEARCH | HEURISTIC EVALUATION & UX AUDIT

To complement user interviews and usability testing, I conducted a heuristic evaluation of the platform using Jakob Nielsen’s interaction design principles. The audit identified usability issues across onboarding, navigation, contextual targeting workflows, and optimisation tools. By combining qualitative user feedback with heuristic analysis, I was able to prioritise high-impact UX improvements for future product releases.

RESEARCH | DEFINE & FRAME THE CHALLENGE

STRATEGY | ROADMAP

With product priorities aligned and core opportunities identified, we began translating research insights into scalable product concepts and workflows.

DESIGN

As a distributed team working post-Covid, we established collaborative design processes using Figma, Slack, and Zoom to support rapid ideation, feedback, and iteration. This allowed Product, UX, Engineering, and Data Science teams to work closely together throughout the design lifecycle.

DESIGN | SCALABLE DESIGN SYSTEM

To improve consistency and scalability across the platform, we created a shared Figma design system with reusable UI components, patterns, and design tokens.
The system helped streamline collaboration between UX and Engineering teams, accelerated product delivery, and ensured a more consistent experience across the platform. Components were aligned with a Storybook library, enabling developers to build and implement interfaces more efficiently.

DESIGN | PRODUCT REQUIREMENT DOCUMENT

Projects began with cross-functional workshops between Product, UX, Engineering, and Data Science teams to align on user needs, business goals, and technical feasibility. Using Product Requirement Documents (PRDs), we defined feature scope, user stories, success metrics, and delivery priorities before moving into design and prototyping. The following example explores the redesign process behind a new homepage experience.

DESIGN | WIREFRAME EXPLORATION

Using the requirements and user flows defined in the PRD, I explored multiple homepage concepts and interaction patterns through rapid wireframing in Figma using components from the design system.

The goal was to quickly validate layout structures, content hierarchy, navigation behaviour, and feature discoverability before moving into high-fidelity design. Selected concepts were reviewed collaboratively with Product and UX leadership to gather feedback and align on a preferred direction.

DESIGN | ITERATION & REFINEMENT

Following stakeholder and UX feedback, the preferred concept was refined and iterated into a more structured homepage experience.

DESIGN | STAKEHOLDER REVIEW

Interactive prototypes were presented to stakeholders and the wider Product team to validate workflows, gather feedback, and align on the final direction before development.

DESIGN | USABILITY TESTING

Feedback from stakeholders and Product reviews was incorporated into the designs before interactive prototypes were tested with account managers. Insights from these sessions informed final refinements ahead of product sign-off and development handover.

IMPLEMENTATION

Successful delivery relied on close collaboration between Design and Engineering. Alongside final UI designs, responsive layouts, validation states, loading states, and edge cases were fully documented to support implementation.

IMPLEMENTION | DESIGN HANDOFF

Final UI designs, interaction states, and supporting documentation were handed over to Engineering through Figma and collaborative review sessions. Features were then broken down into development tickets, prioritised into sprint cycles, and implemented incrementally alongside ongoing design support. 

IMPLEMENTATION 

Before release, new features and updates were validated through UAT sessions involving Product, Engineering, and internal stakeholders. Issues identified during testing were logged, prioritised, and resolved before deployment to production.

IMPLEMENTATION | USER ACCEPTANCE TESTING

Sessions were conducted at the end of each sprint with internal stakeholders and account managers to validate releases before deployment.

IMPLEMENTATION | RELEASED TO PRODUCTION

After iterative testing and refinement, the final experience was successfully launched to production.

ANALYSIS 

Continuous user research and iterative optimisation informed ongoing platform improvements. These enhancements delivered measurable gains in engagement, conversion and campaign performance across multiple industry sectors.

ANALYSIS | PLATFORM RESULTS 

The platform delivered measurable gains in revenue, engagement and conversion across luxury, automotive and hospitality sectors. These results demonstrated the value of combining user research, continuous optimisation and industry-specific personalisation within the 4D platform.