Search within the Optum Now.com platform was fragmented and limited to just articles and products. As users increasingly sought tailored answers, our existing search failed to surface meaningful, personalized results. The experience lacked context-awareness, scalability, and intelligence.

I set out to reimagine & develop an AI-powered search experience driven by:

• Relevance-as-a-Service (RaaS): Dynamically rank and prioritize search modules based on user intent, authentication state, and historical data.
• Retrieval-Augmented Generation (RAG): Serve short AI-generated answers for high-confidence educational queries, then follow with deeper modules.
• Smart Modular SERPs: Each search result page is built from flexible, reusable modules: e.g. content cards, product carousels, dynamix refinement tools, Find Care Carousels, etc – all chosen and ordered by an intent-aware engine.
• Inline Predictive Search: Actively suggests carousels, FAQs, and quick actions as the user types—no need to hit 'Enter'.

• Context-aware hierarchy: "Headache" triggers clinical info + care; "Optum Perks" triggers pricing tools and Rx savings.
• Predictive intelligence: Typing patterns and keyword recognition surface fast actions like "Book visit" or "Refill Rx".
• Authentication-aware personalization: Logged-in users see benefits-based modules; guests see generic suggestions with sign-in prompts.
• Search as destination: SERP becomes a launchpad—educational guide, care connector, and task completer.

• SERPs tailored to intent types: care-seeking, education, product shopping, account tasks.
• Inline suggestions powered by 0PD (zero-party data) like saved doctors, allergies, or health conditions.
• Modular design accelerates development and A/B testing of layout variations.

Outcomes (In Progress, so these are goals):

• Reduced zero-click results or drop-off rates on search pages
• Higher click-through on AI-powered quick actions
• Increased iterative search refinement due to fast acting RAG personalization

Future State:

• Search becomes the primary homepage experience & evolves with user patterns

• Introduce a “Health Guide” persona to narrate or walk users through results
• Expand use of AI-generated modules for symptom queries, plan comparison, and family management

This system transforms search from a generic lookup tool into a dynamic, health-guided experience; one that learns and adapts to each user’s unique needs.

‍