TeachFX AI Instructional-Feedback Coach — Utah Statewide RCT and Ogden City Schools Scale-Up

TeachFX is an audio-only artificial-intelligence platform that records classroom lessons and analyses teachers' talk patterns, then delivers automated, privacy-first feedback by email on metrics such as talk ratio, wait time, and — most notably — teachers' use of 'focusing questions': open-ended prompts that press students to explain and reflect on their reasoning.
Between October 2022 and March 2023, a pre-registered randomised controlled trial run with the State of Utah and TeachFX tested this automated feedback on 224 Utah mathematics and science teachers. Teachers opened the feedback emails 53–65% of the time, and the treatment group asked 4.61 more focusing questions per hour than the control group — a statistically significant 20% increase (p<0.01). The study, led by Stanford's Dorottya Demszky with Jing Liu, Heather Hill, Shyamoli Sanghi and Ariel Chung, was published in Computers & Education in 2024.
Building on that evidence base, the Ogden City School District Board of Education voted on 17 April 2026 to approve a three-year, $216,000 pilot bringing TeachFX to four schools, with unlimited licences, voluntary teacher participation, onboarding workshops and a data-privacy amendment required before rollout. District leaders describe it as offering teachers 'a coach in an app'; the district has committed to reporting usage and outcome metrics to the board during and after the pilot, but as of this writing no Ogden-specific outcome data has yet been published.

Read the full analysis: https://edworkingpapers.com/ai23-875