A Conversation at EdTech Insiders AI + EDU Conference

TeachFX is using AI to help superpower teaching and instructional coaching throughout K-12. We believe that there’s almost nothing more powerful than a great teacher in the classroom, that every student deserves great teaching, and every teacher deserves the support to be the teacher that they want to be.

In that spirit, TeachFX CEO and co-founder Jamie Poskin recently participated in a panel discussion at the EdTech Insiders event, AI + EDU Conference (check out all of the sessions here).

Alongside similarly mission-drive edtech founders including Dr. Jim Wagstaff of Noodle Factory and Maya Bialik of QuestionWell, Jamie explored the promise of AI to help superpower the impact educators can have on student learning.

How do we better support teachers, in a scalable way? 

Jamie founded TeachFX in response to his own experience in the classroom, teaching math and English in Harlem. Driven by his own commitment to equity, Jamie entered the classroom because he sees teachers as the main leverage point for change both within education and our society as a whole. 

In all of his years in the classroom, Jamie was never once observed by another educator. As a teacher early in his career, he craved feedback that he didn’t receive. He was sure there had to be a better way.

And during his work at Stanford's Graduate School of Education, this question – how can we better support teachers, in a scalable way – would spark the start of TeachFX. 

TeachFX is a way for teachers to record audio of their class. The app uses voice AI to analyze that audio in order to provide the teacher with positive feedback and support for their teaching practice. 

The AI is looking at things like the questions a teacher asks, and how you're using certain high-leverage teaching practices like wait time (also called think time) or uptake of student contributions. We show teachers these positive moments in their practice so that they can do more and more of that in the classroom.

It's very similar to what a great instructional coach, principal, or assistant principal might do if they were sitting in the back of your classroom. But, it's all done automatically. 

How can AI help automate routine tasks for educators? 

Teachers’ days are filled with so many routine and repetitive tasks. When AI can do in minutes what took hours, teachers are freed up to dig deeper into the things that brought them into the classroom in the first place – activating students’ thinking and their love of learning.

For instance, QuestionWell saves teachers time by turning the content they’ll teach into questions that spark active learning. What used to take a lot of effort and time for teachers is done so quickly by AI, and that frees teachers to dig deeper, to make those questions better, higher level. Because, as co-founder Maya Bialik describes it, “Memory is the residue of thought.”

At TeachFX, that time-saving focus is on both teachers and instructional leaders.

instructional leaders, much like teachers, are very crunched for time. 

Jamie saw the many reasons for never having been observed during his time in the classroom. Principals and instructional coaches are pulled in a million directions at once. And if you do make it in to do a classroom observation, you’re sitting in the back of a classroom, collecting objective data that you can present to a teacher to reflect on. You want to share feedback that creates space for the teacher to come to their own conclusions. 

But, the way instructional leaders have done this to date has been incredibly time-consuming, laborious, and still can be incomplete. You’re trying to jot down every question a teacher asks, or time how much student talk is happing. You’re not there to keep track of how many a questions a teacher asked, how much the students talk, and so on. You’re there to see the reality of teaching and learning in the classroom, to connect with that teacher, and to be ready to effectively coach that teacher along their professional journey. 

Why sit in that classroom as a timekeeper? AI can do that for you. 

With AI, when you’re in the classroom, you’re focused on what you’re seeing and letting the AI gather those data points for you. 

And, without even being in the classroom, instructional leaders have a way to still provide that same kind of support to their teachers. Because teachers can hit record whenever they want, and gather their own feedback.

A lot of the research on efficacious professional learning shows how time-intensive it is. It shouldn’t take 40-plus hours of teacher time to pursue professional development. By using AI to provide teachers with actionable instructional insights, educators can do in minutes what used to take so many hours.

How does the teacher’s role evolve amid this explosion of AI tools?

At TeachFX, we believe that nothing can replace the power of a great teacher in the classroom. And that includes AI.

We’re not utilizing AI to replace teachers in any way. Instead, we’re using AI to help teachers do better what only humans can do.

The teacher's job is not to deposit information into students' heads. It’s not just about the transfer of information. The teacher’s job is to engage students in dialogue, to be thought partners, and to challenge them to think about the world in these powerful ways. 

As routine tasks become more and more automated, and the transfer of information becomes easier, the teacher’s role will continue to evolve. Teachers will be able to spend more and more time and energy focused on their core purpose, the higher-order work. And that has the potential to reduce teacher burnout.

What insights can AI analytics provide to enhance teaching practices? And how do we know?

Larry Cuban has written extensively about the black box problem of the classroom. We can adopt new curricula, explore new teaching practices in coaching and professional development, or leverage AI to provide teachers feedback. But the question remains…how do we know? 

How do we know that teaching practices are changing? How do we know we're achieving the outcomes we're looking for?

We tend to know a lot about the lagging indicators – the test scores and other metrics that educators have been focused on. But we haven’t known exactly what’s happening in the classroom.

How can we know we’re actually engaging in continuous cycles of improvement, or enhancing equity, if we’re only focused on these lagging indicators?

With AI, there are insights that we can provide to teachers about their own practice. TeachFX is helping teachers focus in on high-leverage practices to engage students’ thinking, like the number and kinds of questions you're asking. With objective data points, teachers can reflect on their questioning techniques: Am I trying to funnel students to a certain answer, or asking a question that focuses their thinking? Am I asking open-ended questions, and how are students responding? 

And we highlight moments when teachers build on what the students are saying. By reflecting back and building on what a student is saying, you’re showing that you’re listening to them. 

We often hear that educators are focused on classroom management, which is a fraught term, but the research shows that the more opportunities you provide students to engage in classroom conversation, the fewer disruptive behaviors you'll have. It's almost a one-to-one correlation. 

And these practices that increase student talk in a lesson matter for equity in education. The research shows that the students who benefit most from engaging in classroom conversations – especially those from historically marginalized groups – get the least opportunity to do so. 

And we know that student talk is a primary causal driver of student learning. There are six decades of research on this. And yet, students talk very little in classrooms today. And that problem is most acute for our plurilingual learners, students experiencing poverty, and students with disabilities. These groups benefit the most from opportunities to talk through their thinking in class, and yet in our in our system today, they get the fewest opportunities to do so. 

It’s a self re-enforcing cycle that our teachers can reverse, with the insights that our AI is sharing with them. 

For instance, a teacher might realize they’re actually asking really low-level questions to their English-language learners, and higher-level questions of other students. And that insight provides them the space to reflect, and to find a scaffolded way for everybody in the classroom to have these opportunities to talk through their thinking. 

Similarly, Noodle Factory captures a lot of in-conversation metrics for tutors, including insights into how students are engaging with the curriculum. How many questions are you asking, what sort of questions are you asking? Sentiment analysis helps teachers to reengage students in the classroom where needed. 

And Noodle Factory is breaking the formative assessments paradigm by using AI to give students more immediate feedback, without requiring more and more hours of teachers’ time on formative assessments.  

What about AI applied to classroom management?

We have now had a few different large-scale randomized controlled trial studies conducted by independent researchers from Stanford, Harvard, and the University of Maryland, on insights and feedback that are in our app.

Teachers getting this feedback increased their use of high-leverage practices. And it had an impact on student outcomes, which is phenomenal. 

But one of the surprising results of one of these RCTs was that teachers who were getting this feedback actually had higher job satisfaction and morale than the teachers in the control group.

When we ask what impacts burnout, it's almost looking at the problem wrong to look at classroom management first, to say: ‘Oh, I want to have a noise Activity Monitor manage my classroom.’ 

It turns out that helping minimize teacher burnout is more about being an effective instructor. It’s about creating productive conversations, where students are engaged in thinking out loud. 

It's really about reframing that that I think really helps and that's a big part of what we help teachers see and do. And I think a big reason why we found this somewhat surprising result in the in the in the randomized control trial study.

What is one thing that you most hope comes from AI's impact on education?

The promise of AI can be both extremely exciting and a bit overwhelming at the same time.

Are you feeling hopeful about AI’s potential impact on education, and the educators in our classrooms? Drop us a note to share your thoughts!

Administrators & Coaches

Classroom Teachers