For many digital teams, search and recommendations are foundational to the customer experience, yet they’re often built on fragmented, legacy systems that struggle to keep up with rising expectations. As shoppers become more specific and accustomed to personalized experiences, “no results found” is no longer acceptable.

In this session, Dianna Lyngholm, Director of Creative & Website at HalloweenCostumes.com (part of umbrella brand, Fun.com) will share how the online retailer modernized its digital search and discovery experience by rethinking how AI-powered recommendations and search work together. After years of relying on a combination of homegrown search and third-party recommendation tools, the team took a research-led approach to upgrading its stack, evaluating several solutions, A/B testing, and prioritizing relevance, personalization, and long-tail discovery over quick fixes.

The session will unpack what it takes to implement modern, AI-driven discovery on a highly customized platform. From handling cold-start search and seasonal demand to aligning internal teams, our discussion will take a look at the challenges involved in upgrading search and recommendations systems and reveal how Fun.com approached them. Attendees will also hear how better search and recommendation design can unlock measurable CX and revenue gains while maintaining internal trust and control.

Attendees will learn:

• How to modernize search and recommendations to support more personalized digital journeys.
• Lessons from implementing AI-driven discovery on a custom platform, including change management and internal adoption.
• How improving relevance and long-tail search can drive measurable CX improvements and commercial impact.