Universal Search & AI Mode
2026
I initiated and led the redesign of Mediately’s core search and navigation experience across iOS, Android, and Web. By replacing hidden tabs with a smart, priority-based Universal Autocomplete and rebranding our AI assistant into a cohesive "AI Mode," we dramatically simplified the app and made high-value medical tools instantly findable.
+ 72.7% Unique user adoption
+ 10.4% Total search activity is now AI
+ Unified Search
+ AI integration in main flow
Unifying the Clinical Mind
Mediately (medical EU app for doctors) legacy tab-based navigation hid some of our most valuable clinical features. We replaced this friction with a unified, smart search experience that bridges the gap between different medical data types. Whether a doctor needs a local drug leaflet, a complex diagnostic calculator, or an AI-generated treatment suggestion, they are now exactly one search query away from the answer.


Designing for Clinical Speed
In a high-pressure clinical setting, every second counts. Traditional search engines simply match characters, but we engineered our autocomplete as a clinical routing engine.
Instead of overwhelming doctors, the system works quietly behind the scenes:
-
Intent-Based Routing: Typing keywords like "classification" or "icd" instantly surfaces the specific product section rather than forcing the doctor to scroll through pages of irrelevant drug names.
-
Smart Actions: Every autocomplete suggestion maps to exactly one highly predictable action: opening a deep-linked direct item (Drug, Tool, Guideline), running a query-based search, or launching AI Mode.


Engineering for Hospital Realities
Hospital Wi-Fi is notoriously unreliable, often dropping entirely deep inside concrete buildings. We designed defensive UX patterns to ensure doctors are never left without answers:
-
Offline Fallback: If a premium (PRO) user loses internet connection or the autocomplete API fails, the interface gracefully falls back to local database results, displaying a subtle toast: "Showing local results."
-
Behavior-Based AI Shifting:
-
If a doctor types one word, we default to local, super-fast Universal Search.
-
If they type two or more words, the system automatically anticipates a complex clinical question and positions an AI Mode suggestion right at the top.
-


Unifying the AI Brand Experience
As our product matured, we needed to transition from a playful AI assistant ("Elly") to a polished, authoritative clinical workspace ("AI Mode"). I directed this structural and visual rebranding across all user touchpoints.
​
-
Clean Interface, Less Noise: We stripped out redundant "PRO" badges inside the chat, removed programmatic ad placeholders when they weren't needed, and modernized the input fields with a floating layout.
-
Decluttering the Home Screen: We simplified the home view by removing legacy navigation blocks and introduced a persistent "Prescribing" label. This allows doctors to instantly jump straight to critical dosage and indication details without scrolling.


Iterative, Data-Driven Evolution
Search ranking is only as good as the behavior it supports. Working closely with our engineering and analytics squads, we ran rapid feedback loops to constantly fine-tune our result hierarchy. By analyzing real-world clinical search data, we dynamically elevated high-intent, actionable tools over low-priority reference materials. constantly adapting the search engine to mirror the actual urgency of a doctor’s live workflow.

