Interaction models for Indic-first products

This note is a research direction, not a completed study. It outlines questions we believe deserve serious investigation.

Global product patterns often assume typed English, high literacy, and stable connectivity. India’s usage reality frequently assumes voice, code-mixing, interrupted attention, and shared devices.

We are interested in evaluation methods that measure usefulness under dialect variation and noisy environments - not only clean benchmark text.

Related work in national language infrastructure creates opportunity. Product design must still ask what people are trying to accomplish, and what failure costs.