“Zari: Exploratory Search”
Problem
The knowledge about a topic is dispersed across a variety of sources: books, newletters, tweets, podcast clips, and web pages. It cannot be easily ranked algorithmically.
Solution
Zari is a tool to explore the network of related topics and sources for a query, and collect their findings into guides.
zari (n) - a type of gold thread woven into fabrics in Indian clothing
Storyboard
The user starts the journey by searching in the search box. The query can be also be a natural language question.
Semantically relevant results are shown, grouped by their source (e.g. creator, subreddit, etc).
If a particular source seems promising, the user can narrow their search within that source to find more relevant content.
The user can also pivot their search based on semantically relevant terms that are mentioned in the results.
As the user finds relevant snippets of content, they can add it to a “shopping bag” type collection for their search.
The contents of the “shopping bag” can be arranged into a Zari thread and shared so that others can benefit from the user’s curation and research.
Value Prop
Searchers can pivot their queries to understand a topic, in a way that is not supported by traditional search engines.
Other users can benefit from the ‘golden threads’ that searchers before them have curated.
