ANet / TeachFX
Joined 2026
Partners
What are we building?
ANet and TeachFX are partnering to test whether integrating TeachFX’s automated feedback into ANet's curriculum-based professional learning (CBPL) model can improve HQIM-aligned math instruction, strengthen coaching quality, and reduce coaching costs at scale. Alongside the pilot, TeachFX will co-design new coaching feature prototypes and implementation supports with ANet coaches and teachers. While TeachFX focuses on providing teachers access to their own classroom data, these prototypes explore how coaches and teachers might engage with that data together—supporting shared reflection and more targeted coaching conversations, and how doing so might extend the quality and impact of instructional coaching.
What are we learning?
Our overarching learning question is:
Can integrating AI-generated classroom feedback into ANet's coaching model improve teacher practice and coaching quality—and at what cost?
Specifically, we aim to learn:
Teacher practice: How does TeachFX’s automated feedback alongside ANet coaching impact teachers’ use of HQIM-aligned teaching practices?
Teacher efficacy: How does TeachFX’s automated feedback alongside ANet coaching impact teachers’ confidence and self-efficacy in facilitating rigorous math instruction?
Coaching quality and reach: How does TeachFX’s automated feedback data change the nature of coaching conversations? Can it help coaches provide more targeted, evidence-informed feedback without increasing their time burden?
Cost efficiency: Can a TeachFX automated feedback + ANet coaching model meaningfully reduce per-teacher coaching costs while maintaining the quality outcomes associated with ANet's traditional coaching model?
Prototype design: What features and implementation supports do teachers and coaches identify as most likely to help them incorporate automated feedback data into their existing coaching routines and extend the quality and impact of their coaching?
Products
During the first year of this project, TeachFX will prototype and test the following, co-designed with ANet coaches, participating teachers, and external research partners from Harvard/MQI Coaching:
Coaching feature prototypes: New coach-facing features that enable instructional coaches to review teachers’ automated feedback, identify patterns across teacher recordings, and prepare for focused, curriculum-aligned coaching conversations. These features will be iteratively co-designed and co-developed with coaches and teachers throughout the pilot.
“Getting Started with AI-Extended Coaching” Toolkit: step-by-step guidance for school leaders on piloting automated feedback-enhanced coaching models, including participation criteria, staff communication templates, example coaching cycles integrating TeachFX analytics, and facilitator materials (slide decks, one-on-one, and professional learning community coaching protocols).