Hackathon: AI-Powered Search Tag Generator
An AI-powered tool that automates and enhances the creation of search tags while onboarding a brand, improving efficiency, reducing brand creation time, and improving tag quality.
ROLE
Product Designer
TIMELINE
3 Days
COLLABORATION
2 Engineers
Project Management Team
Overview
The AI-Powered Search Tag Generator was developed as part of a hackathon project in June 2025. The goal was to address the pain points in the manual process of adding search tags for new brand partners within an admin platform that uses Algolia for search.
Impact
Before
3X
New brand creation time
Reduction
80%
Manual input errors
Reduction
After
Goal
Accelerate the organization onboarding process
Increase brand discoverability on the platform
Enable a seamless AI-assisted workflow
User Research
3
Project Managers
120
Minutes
3
Days
-
Conducted interviews with the Product Management team to understand the current workflow and pain points.
-
Found that despite using Algolia, search tags were not automatically generated and required heavy manual effort.
-
Identified that the tagging step was the most effort-heavy, making org creation slower and inconsistent.
Quick Prototyping
For the hackathon, I used AI design tools like Bolt.ai, v0.dev, and Lovable to rapidly prototype the search tag generator interface. These tools helped accelerate design iterations within the limited time, enabling quick exploration of AI-assisted UX concepts and seamless integration with the existing platform, resulting in a 50% faster prototyping.
Problem
Manual tagging was identified as a major bottleneck in the onboarding process for creating organizations (orgs) on the platform. Each org required 10 to 20 search tags, which users had to generate manually by browsing brand websites and copy-pasting keywords. This resulted in:
-
Time-consuming and repetitive workflows
-
Inconsistent tag quality is dependent on human judgment
-
Risk of poor brand discoverability due to missing or inappropriate tags
Manually search and find brand tags
Solution
The solution automates search tag generation by using AI to extract keywords from brand website content and integrates these suggestions directly into the org creation workflow, streamlining the process. This reduces manual effort, ensures consistent tag quality, and improves brand discoverability.
BEFORE
AFTER
Enter the brand’s home page to generate tags
Space to enter tags manually if required
Disclaimer of how the tags are generated
CTA to generate more tags
CTA to select all the generated tags
Confidence score to show the impact of the tag
Tooltip info on why the tag was generated
CTA to apply the selected tags
CTA to deselect all the tags
Action to clear all the selected tags
Selected tags in pill format to have better visibility over selections and action to remove specific tags
Technical Architecture
Collaborated early with engineers to understand backend capabilities and constraints to build this functionality. Here are some of the details I found about the technical requirements.
Web scraper extracts content from brand URLs
AI engine (GPT-4 or LLAMA2) generates tags
Key Learnings
Clear explanation of AI-generated results builds trust
Early feedback from stakeholders improves design direction
Automation can drastically reduce workflow time