Letter from the Founder: TeachFX raises $10M to create more meaningful and equitable classroom discourse
Six years ago, I started building this nifty little app for teachers. The app was pretty simple: I’d record my class, and the app, which I called “TeachFX,” would send me back insights about my teaching and the learning dynamics in my classroom
Was I listening to my students? Was I asking open-ended questions?
What did my lesson design look like? What was the balance of teacher vs. student talk?
Which students were getting to do the heavy lifting? Was that equitable?
Why does this matter?
These are some of the things we can do that research says promote better student learning outcomes. But there’s more to it than that -- dialogue is how we encounter and relate to other human beings; it’s how we shape our identities and become our full selves; it’s how we make meaning in the world. I wanted my classroom to foster deeper understanding of ourselves and others. That isn’t possible without dialogue.
My idea was the app would do all this analysis automatically with AI. Eventually. But at the time I still hadn’t met my genius cofounder Berk Coker, and as a former teacher with fairly limited technical skills myself, I was labeling data manually -- the “Turk” hiding inside the supposedly mechanical chessboard.
I am incredibly excited to announce today that we’ve brought $10M of new investment into TeachFX. Today, the TeachFX app is real -- doing everything I’d hoped it would back in 2016, and more, using cutting-edge voice AI we’ve developed that analyzes teaching practices such as levels of questioning, use of wait time 1 and 2, and uptake of student contributions.
We’re now in the hands of over 8,500 teachers across every state in the country. Our partnerships with schools and districts are expanding rapidly -- 9x growth in the last year alone. And in our partner districts, license utilization of the TeachFX app is over 100% (implementations of TeachFX tend to spread organically to teachers beyond the initial cohort or school).
The Vision
I started TeachFX because I believed that a skilled teacher generating passionate discourse in their classroom is the key to advancing equity and unlocking deeper learning for students. That’s what I always wanted as a teacher! But it’s tough. Teaching is such a demanding job, and I rarely had the chance to reflect on my instructional practice. Some days my students were engaged, some not. Professional growth came slowly for me. Too slowly -- I was out of the classroom before I’d gotten good enough at the job to feel that sense of mastery we all need to be fulfilled in our work.
Then, while studying at Stanford’s Graduate School of Education I met amazing mentors like Rachel Lotan and David Brazer and encountered research that reminded me of why I started teaching in the first place. The ideas that stuck with me most now serve as the key research principles that drive our work at TeachFX:
1. Student Voice:
Students need to speak in order to learn. The research here is unequivocal and spans five decades (Howe et al., 2019, Johnson, 2019, Murphy et al., 2018, Holthuis et al., 2014, Michaels et al., 2008, Lotan, 2003, Mortimer & Scott, 2003, Roth & Lee, 2002, Bianchini, 1998, Cohen & Lotan, 1997, Lotan, 1995, Driver at al., 1994, Leechor et al., 1989, Vygotsky, 1978).
The relationship is causal: give a student more opportunities to engage in discourse in the classroom, and they will learn more -- by any measure. As many educators put it: “The person who is doing the talking is doing the learning.”
So, what does this mean for us teachers? Getting students engaged in meaningful and equitable discourse is no easy task. Teachers might be working on using high-leverage teaching practices; they might be implementing a new curriculum. All our efforts have the same end goal: deeper student engagement with the material -- particularly through discourse! At TeachFX, we believe it’s essential to measure that.
2. Equity:
Teachers talk 70-80% of class time (Hattie, 2012 ). The remaining ~25% includes activities like quizzes, worksheets, transitions. This leaves precious little time for students to talk.
Take this startling fact: at TeachFX we’ve analyzed over 100,000 hours of audio from K-12 classrooms across the country -- the largest-ever study of student discourse -- and found the average student spoke only 7 seconds per hour of instruction. But this is only part of the story.
Historically marginalized populations -- students from low-income families, students with disabilities, English learners -- benefit the most from speaking in class but get far fewer opportunities to do so than their peers (Fisher et al., 2008). We saw these dynamics in our Equitable Classrooms study, with Black and Latino students speaking less than half as much as their peers in many districts.
The reality of K-12 education in the US today is that the most powerful opportunities to learn -- through interacting -- are inequitably distributed. Despite decades of efforts, inequities persist. But when we change who does the talking, we change who does the learning. With our partner districts like Anaheim and Detroit, we are doing just that.
3. Feedback:
In order to improve in anything, we need feedback. Teachers are no exception; we need frequent feedback that is personalized to our practice in order to grow (Brazer et al., 2018).
But teachers rarely get authentic or personalized feedback. Teachers have long been supported through “professional development” (PD), but it isn’t working. PD is mostly loathed by teachers (Gates Foundation, 2016) because it’s usually a one-size-fits-all, sit-and-get affair -- totally disconnected from our actual work teaching in the classroom. On top of that, traditional PD requires far too much time of teachers (50+ hours!) to make any impact on teacher behavior (Darling-Hammond, et al., 2009, French 1997).
TeachFX changes all that. With the push of a button, teachers generate rich, classroom-level data whenever we want it. Reflecting on that feedback for less than 5 minutes reliably produces positive change in teacher practice. This unlocks a new way to lead instruction and professional learning -- with feedback on what truly matters, so easy to produce that it can be incorporated into every aspect of the job.
What prompted me to seek to expand our impact was this: Unlike other approaches to improving student outcomes by changing teacher practice, we have evidence that TeachFX actually works! Teachers using our product improve the way they teach, in ways we have measured directly: they ask more open-ended questions, use wait time more often, and use other teaching strategies that produce more meaningful and equitable student discourse in the classroom. Over this past school year, for example, teachers using TeachFX increased their student talk by an average of 40%.
We know that teachers are the highest-leverage factor in the entire education system. With TeachFX, we saw in our data that we had discovered an extremely low-cost, high-impact way to transform education for the better.
Then Prof. Dora Demszky, one of our collaborators at Stanford, told me that her research group from Stanford, Harvard, and the University of Maryland had run a 5-week randomized controlled trial experiment with 1,136 teachers studying the impact of the insight in our app on teacher uptake of student contributions. The results knocked me out of my chair:
Teachers who were randomly selected to receive the feedback improved their use of uptake by 24% compared to the control group.
Students of those teachers reported higher course satisfaction, rated their teachers higher, completed more assignments, and had higher academic achievement than students of teachers in the control group.
Not only were we seeing this incredible efficacy in our own data, but the results held in a large-scale RCT run by the top researchers in the world. No other PD intervention has anything close to this kind of research validation -- our intervention is proven to change not just teacher practice but actual student outcomes as well.
And this intervention doesn’t cost 50+ hours of teacher training time; the teacher just spends a couple minutes reviewing their insight data. With this validation, we at TeachFX feel a responsibility and urgency to get this into the hands of every teacher in the world.
With TeachFX, every teacher in the world can have access to the feedback and self-reflection that changes their practice and improves student outcomes. To usher in this new era of teaching and learning, it will take change. New ways of thinking. But today we are stuck in old conversations - conversations we’ve become far too comfortable having - with little action taken to change them. A handful of these conversations:
Teachers doing most of the talking instead of students.
Teachers receiving little to no personalized support.
Deeply entrenched equity challenges are robbing students of the opportunities they deserve.
TeachFX is on a mission to change these conversations:
We want to change the conversation in the classroom.
Our app helps teachers understand who is speaking and why -- what “talk moves” are teachers making to promote discourse? -- so that conversations in the classroom are more equitable and more student-led. 🙋🏽♂️🙋🏼♀️🙋🏿♀️
We want to change the conversation about “classroom management.”
Here’s a phrase that makes my skin crawl: “classroom management.” Like these human beings whose learning we have the honor of facilitating need to be “managed” into quiet, non-disruptive compliance. But how compliant would you be if you were forced to sit in a classroom where you only got to speak for 7 seconds an hour? Would you rebel by tuning out or acting up? 😈
Guess what teacher behavior corresponds most strongly to fewer disruptive behaviors in the classroom? It’s not “quiet coyote” or positive behavior narration. It’s the number of questions the teacher asks; every question a teacher poses leads to one fewer disruption (Sutherland et al., 2003).
Instead of managing kids into compliance, let’s engage them through discourse.
We also know that our current classroom management practices lead to students of color being disciplined disproportionately to their white peers (Gordon, 2018). Imagine if we saw students’ eagerness to speak not as a threat to a “well-managed” classroom but as curiosity and excitement to engage with the material?
We want to change the conversation about student mental health.
We are in the midst of a mental health crisis in our schools. Social emotional learning (“SEL”) has appropriately been a focus for many whose students are still emerging from months of social isolation. Student discourse is the sine qua non for social and emotional learning to take place. Start there.
As our data showed, students were only getting the chance to speak for a measly 7 seconds per hour of class. Engaging in dialogue with students is the most powerful way to build a sense of belonging and community in the classroom. TeachFX provides insights to teachers about when we affirm what we hear students saying, and when we are posing “authentic” questions that we as the teacher truly do not know the answer to, like “how did you get to that answer?” or “which character did you identify with the most, and why?”
These are ways teachers make students feel welcomed, to feel a part of a community, to feel that their ideas matter and they deserve to be listened to.
We want to change the conversation about supporting teachers.
These last few years have been so challenging for educators; it’s not surprising that teachers are leaving the profession in droves. There are dozens of ways schools and districts are using TeachFX to take work off teachers’ plates -- from eliminating the need to collect evidence manually for teacher collaborative team meetings, to replacing or streamlining PD protocols and requirements.
But it goes beyond that. Teachers are leaving not only because of how much they’re asked to do. Educators are frustrated about the de-professionalization of teaching, about “being micromanaged by rigid curricula that turn them into little more than data collectors” (Greenblat, 2022). This makes teachers feel less professional and less fulfilled.
Instead, teachers need to be treated like the professionals they are: professionals capable of leading their own learning, of reflecting and growing. Teachers, like all of us, need to be constantly learning in order to be fulfilled in their jobs -- to feel a continued sense of mastery in our profession in order to thrive. Most importantly, teachers need to connect to our purpose, to why we became teachers in the first place. Focusing on the learning dynamics of my classroom allows me to do that.
TeachFX enables all these conditions. We give teachers autonomy -- no one can see a teacher’s TeachFX data but themselves. We put teachers in the driver’s seat of their own learning for continuous improvement and ongoing mastery in the profession. Critically, our data shines a spotlight on what really matters to their success and fulfillment as educators: creating an engaging classroom.
We want to change the conversation about professional development.
71% of teachers are dissatisfied with the PD they’re receiving despite schools spending $18 billion on that PD (Gates Foundation, 2016). The same study found that teachers want PD that is “interactive” and “sustained over time.” It must “treat teachers like professionals.” Most importantly: “it has to be personalized.”
The reality stands in sharp contrast: We sit through a PD, then it’s over, with no feedback or follow-up. PDs happen to us a few times a year -- the same thing for every teacher without any personalization that connects to our daily practice or specific context. Some teachers experience “walkthroughs” where an administrator might observe 5 or 10 minutes of their class. Most teachers aren’t observed at all.
And many teachers don’t want to be observed! The relationship between admin and faculty is often fraught. So administrators perform evaluations once every few years. But these are almost never about teacher growth. Some teachers have an instructional coach, but the vast majority don’t, and even those who do might be observed a couple times a year at best (ibid.). Nonprofits such as TNTP and Digital Promise report the only effective PD requires frequent, objective feedback…but very few teachers get that.
The bottom line: K-12 education has a human capital challenge -- it’s costly to truly support a teacher, to sit in the back of their classroom and observe, to collect data and take notes, to prepare those notes for a conversation with a teacher. Because it’s costly, it happens very infrequently, and teacher supports like traditional PD are designed around this constraint -- the lack of that most precious resource, time.
K-12 schools can now take advantage of labor-saving technology that makes possible ongoing, job-embedded support for all teachers. 🎉
We want to change the conversation about data in K-12.
Right now, we still live in the shadow of No Child Left Behind, which required states to develop and administer assessments to all students in order to receive federal school funding. The law gave rise to a gigantic corporate ecosystem -- the assessment industry, or what some refer to as the Testing- or Assessment-Industrial Complex -- set up to profit from this de facto mandate.
Teachers today spend 40+ hours a year working in teams reviewing student answers to multiple-choice questions on high-stakes tests (Hill, 2020). Many teachers spend more time working with this data than preparing for lessons. And yet these tests are of racist origin of questionable value.
Another problem: knowing that half my students struggled with algebraic manipulations or with grade-level vocabulary doesn’t tell me how to teach differently. So, I teach the same way, but spend extra time on different topics depending on what the assessment data tells me. It’s rearranging deck chairs on the Titanic. Unsurprisingly, the research shows that this massive time investment makes no impact on student outcomes (Barshay, 2022, Hill, 2020).
So, what data should teachers be studying? The research is clear here too: data from the classroom, that is, data about their own instructional practice, ideally in collaboration with their colleagues (Brazer et al., 2018). Only by reflecting on what I did as a teacher and the impact that had on students -- the learning dynamics of my class -- can I truly understand what I could do differently next time.
Today, however, the learning dynamics inside the classroom are not measured at all. Why not? Well, because until now, collecting this classroom-level data has been extremely time-consuming and expensive -- you’d need an expert in pedagogy sitting in the back of every teacher’s classroom on a regular basis.
With TeachFX in the hands of every teacher, we are no longer imprisoned by the narrow view of students as numbers, as scores on high-stakes tests. If we free ourselves from the mindset that the only “data” we have in education is test scores, we open up a new world of teacher learning and growth.
We want to change the conversation about equity in education.
By far the most important conversation we seek to change, and the reason for changing all the others, is to provide all students the rich opportunities to learn they deserve.
-
Because of the primacy of student assessment data, it is common for teachers to be “introduced” to students through their assessment results. This shapes teachers’ expectations about student readiness and capacity for learning. Those expectations can be reinforced by the historical patterns in academic performance for specific demographics or student groups. These achievement “patterns are so regular and so predictable that,” much-revered professor and teacher educator Gloria Ladson-Billings argues, “we have come to expect them” (Ladson-Billings, 2013).
Our expectations for students, research shows us, is one of the most powerful forces that impacts learning outcomes. The most famous example might be the “Pygmalion” study conducted in 1968 by Robert Rosenthal, a Harvard psychologist, and Lenore Jacobsen, a school principal in San Francisco (Jacobsen & Rosenthal, 1968). The researchers administered “pre-intelligence” tests to all students at an elementary campus in California. They then proceeded to inform teachers which of the students showed the “most intellectual promise” according to the test. Eight months later, the students were all assessed again, and those singled out by the researchers for intellectual promise showed significantly higher gains in academic performance.
Here’s the twist: Those supposedly “most promising” students were selected by Jacobsen and Rosenthal entirely at random.
Teacher expectations made a stronger impact on student achievement than established measures of “ability.”
The research I cite above from Lotan and others shows that the “gaps” we see in student outcomes across race, income, language learner or ability status are a product of who has a chance to engage in meaningful dialogue in the classroom, and who doesn’t. The gaps between students of different races, for example, disappear when students are given equal opportunities to speak, even for supposedly “fixed” measures of intelligence such as IQ (Lotan, 2022). We must change the conversations in our classrooms to achieve great equity.
Join us in changing the conversation!
When you start a company you have a lot of things to prove. Does anyone actually want your product? Does your product really do what it says on the label? Most importantly for an education company like ours: Does your product advance equity and student learning in meaningful ways?? We spent years honing our ML algorithms, testing which methods of feedback were most helpful to teachers in reflecting on and changing their practice, and convincing our first school and district partners to join us on this crazy mission to transform teacher professional learning and instructional leadership. Today we are taking a major step forward.
The investment comes from incredible institutions who believe in our mission and vision: Reach Capital, an impact fund focused on education who led the round, along with Osage Venture Partners, Pearson Ventures, Brighteye Ventures, and a host of angels (shoutout to Ulu Ventures and Newfund Capital who, along with Reach and Brighteye, invested in our seed round). This capital will help us scale our impact in K-12 education, expanding into live-online learning, continue to develop our cutting-edge voice AI technology, and help us to build a research consortium to advance the field.
Are you a fellow traveler on this journey?
If you’re reading this and you are ready to join in our mission to support teachers in promoting more meaningful and equitable classroom dialogue, here are a couple ways you can get involved:
Partner with us! TeachFX works with K-12 schools and districts, colleges, and tutoring providers to transform professional learning and instructional leadership to harness the power of automated teacher feedback. If you’re an educator and think your school could benefit, drop me a line at jamie@teachfx.com.
Help spread the word! As a parent, I want my three young kids to be in classrooms filled with student voice.
(Thank you to Victoria Dye and the wonderful teachers in Sequoia Union, where my children will attend, for launching our partnership there last year.)
Parents are the most powerful force in the education system. Ask your local school or district about TeachFX!
Join our team! Openings are posted on our careers page. We are always looking for passionate individuals who want to work in service of our mission.