In sports media, timing is critical. However, editors face the time-consuming task of manually searching through hours of game footage to find the perfect highlights. To address this, the Ads Creative Engine is an AI-powered search and discovery tool that transforms how editors find, manage, and use sports video assets for global ad campaigns.
The Challenge
Weekly game footage overwhelmed creative teams, who received ‘pre-broken’ clips without a centralized, searchable way to find specific moments. To meet this need, the project aimed to:
- Reduce Search Time: Move away from manual browsing to a fast, transparent search tool.
- The client initially did not realize how Generative AI could solve their specific bottlenecks.
- Scale Content Production: Enable a single editor to efficiently produce up to eight multi-platform video sets (Meta, YouTube, TikTok, Snapchat) per week.
To tackle these challenges, I served as the Lead Designer and Product Strategist, bridging the gap between AI’s technical potential and creative needs. I drove the product vision from zero to a high-fidelity prototype under a tight timeline.
- I led workshops and interviewed users to uncover the real needs of editors and marketing staff.
- End-to-End Design: As the sole designer, I was responsible for the entire UX/UI lifecycle, including user workflows, wireframes, and final mockups.
- I educated stakeholders on the possibilities of GenAI and transformed technical options into a finalized product roadmap.
The Process
The project advanced in clear phases to deliver immediate results and support future growth:
- Phase 1 (The Core): Launch a searchable library with precise clip descriptions, transparent metadata, and bulk download tools.
- Phase 2 (The Ad Concept): Build a generative tool for editors to streamline ad plans and automate creative roadmaps. With the project phases outlined, focus shifted to discovery and user research, which informed critical design choices.
Discovery & User Research
- I interviewed Social Media Editors and Sports Marketing teams to map the current Clips Distribution Process.
- Identifying High-Value Content: Editors needed a way to instantly find “exciting” footage: touchdowns, interceptions, crowd celebrations, and player reactions.
- Metadata Importance: Date, season, team, and player number are essential search filters.
- Pain Point Analysis: Feedback on existing Proofs-of-Concept (PoC) highlighted low accuracy and the delivery of “useless” one-second clips as major blockers.
Ideation & AI Roadmapping
- Once the needs were clear, I facilitated ideation sessions to show possible functionalities.
- Natural Language Queries: We built a system that supports natural-language search, enabling editors to type “Messi goal with crowd reaction” and receive instant, accurate results.
- Creative Expansion: Proposed future features like “Audio Search” (announcer name calls) and AI-generated social media captions based on clip content
- Ad Assistant Flow: I defined a new process for ideation before editing an ad to optimize the workflow.
UX/UI & Workflow Design
- I designed a clean, “search-first” interface that minimized UI “real estate” to prioritize video results.
- To further support seamless workflows, I designed a visualization tool to track video processing and indexing in real time within the upload pipeline, ensuring users remain informed at every stage.
- Multi-Level Workflow: Created a comprehensive user architecture (Level 0 to Level 3) covering everything from Sign-In to Bulk Download and AI Ad Planning.
- Ad Concept Flow: I defined this creative process to enable users to maximize value from the platform.
Outcomes
- Designing for Trust: Addressed potential AI failures (latency, bias, model drift) by designing clear user feedback mechanisms.
- Simplifying Complexity: Translated sophisticated backend processes (Vector Embeddings, API Gateways) into a simple, intuitive search bar for non-technical creatives.
- Leadership through Design: Acted as both the Product and Design lead to move a client from “AI curiosity” to a tangible, ROI-driven marketing tool.