Natural-Language School Search

My role: AI Product Architect & Developer

Starting problem

Parents had to work through multiple dropdowns and checkboxes to express what they wanted from a school. The education platform already had search functionality, but using its filters required parents to translate their needs into the platform's criteria themselves.

My contribution

I owned the feature from the initial client discussion and problem definition through design, development and testing. I integrated natural-language processing with the existing search: a parent's written requirements became curriculum, location and age-group criteria. I tracked actual customer requests and the resulting outputs, then manually checked whether the interpretation matched what the parent intended.

Illustrative request and intended interpretation Request: "I'm looking for a British-curriculum school in Dubai for my eight-year-old." Intended interpretation: curriculum — British; location — Dubai; age — 8, using the platform's existing age-group criterion. This is a written example created to explain the workflow, not a recorded customer request or a validated system output.

Key decisions

• I added natural-language input to the existing search workflow, preserving its structured criteria and search functionality.

• I mapped written requirements to curriculum, location and age-group criteria rather than leaving the output as an unrestricted text response.

• I checked actual requests and their outputs manually against the intended interpretation as part of testing.

Delivered capability and status

The feature was tested and deployed to production within the existing education platform. My delivered responsibility covered discovery, feature design, development and testing. Manual review of customer requests and outputs is the supported validation method.

Image note Workflow illustration of natural-language school search; it is not a screenshot or recorded customer request.