A product story
From 7 Days to Hours: Building an NLP Search Engine for Educational Content
Enabling instant content discovery across thousands of nested educational objects using conversational AI.
Reduction in Course Build Time
01 / The problem
The situation
before the shift.
Content developers build courses to meet specific state education standards — for example, Texas might require 50% of a maths course focused on Algebra 1. To fill gaps, developers needed to find relevant lessons, questions, and items across thousands of objects nested several levels deep. The process relied on manually navigating hierarchies, separate spreadsheets, and tribal knowledge from a handful of experts. A complex course build could take 1 to 7 days, affecting roughly 80% of content developers.
02 / The hypothesis
Identified the problem firsthand, validated through ride-alongs with power users, and quantified the impact using Pendo analytics to build the business case.
03 / The approach
- Discovery: Identified the problem firsthand through direct observation and power user ride-alongs.
- Validation: Quantified the impact using Pendo analytics to build a compelling business case.
- Prototyping: Led discovery with AI-assisted low-fidelity prototyping to validate the concept before engineering commitment.
04 / Execution
- Delivered a conversational search tool using NLP technology.
- Enabled natural language queries against the CMS for intuitive content discovery.
- Implemented search across thousands of nested educational objects and multiple hierarchy levels.
- Measured the impact through Pendo analytics and stakeholder feedback from published courses.
05 / The result
50%Reduction in Course Build Time
The implemented feature is delivering a 50% reduction in course build time, evidenced through Pendo analytics and stakeholder feedback from teams working on published courses.
- 50% reduction in course build time after implementation.
- Impact evidenced through Pendo analytics.
- Positive stakeholder feedback from published courses.
- Addresses pain point affecting 80% of content developers.
- Reduced dependency on tribal knowledge and manual navigation.