Making the trust between students and educators measurable — natural language processing, $1.25M to Sown To Grow (SBIR Phase II)

1Measuring something hard to measure

The quality of a relationship between a student and a teacher is said to predict outcomes well, yet unlike a test score it cannot be read off directly. The abstract notes that many schools lack practical tools to assess or support these relationships in real time at scale.

What the project takes on is that hard-to-measure thing, reading the state of a relationship from what students write in their reflections through natural language processing. Quantification is not the goal in itself; the stated aims are letting teachers respond more meaningfully and letting administrators see where support is needed.

2The competitive advantage is placed in the dataset

The abstract states plainly that the competitive advantage lies in a proprietary dataset of millions of reflections. What is named as the strength is the quantity and quality of material in hand rather than the sophistication of a model.

Student writing is material that cannot readily be bought in. It accumulates in a tool through long use, and a later entrant cannot easily assemble the same. The difference forms along an axis separate from technical merit.

3Companies hold 13 of the 120 awards

Among the 120 NSF awards this site holds as of 2026-09-01, recipients break down as 104 universities and the like, 13 companies and three independent research institutes. Awards to companies come to a little over a tenth of the whole.

The obligated amount is exactly $1,250,000, matching the estimated total. SBIR Phase II amounts tend to gather near a ceiling, and several records this site holds carry the same figure. The period runs two years, from 2026-08-01 to 2028-07-31, shorter than the three to five years common on university awards.

4Thirteen of 120 awards go to firms

Sorting the 120 NSF awards this site holds as of 2026-09-01 by the kind of recipient gives 104 to universities and similar bodies, 13 to companies and three to independent research institutes.

Awards to companies come to rather more than one in ten. Research funding is a university instrument by design, yet frames aimed at small firms, such as SBIR, sit in the same list, so the recipients are not of one kind. This record belongs to that one in ten.