People often ask us why we decided to make Speak Agent. As co-founder and CTO, I wrote the first line of software code over three years ago. But the need has been around for much longer. Over the past decade the percentage of English Language Learners (ELLs) achieving a Proficient level on the National Assessment of Educational Progress has actually dropped. It's now roughly 1 in 20 for reading. In fact, 70% of ELLs do not even achieve Basic, the lowest level. This, in turn, causes nearly 40% of ELLs to fail to graduate high school in 4 years. Many students who speak English as their primary language also have limited proficiency because of gaps in academic language that appear in early childhood.
According to the Joan Ganz Cooney Center, the number of "book words" known at age 5 is the single greatest predictor of K-12 success. Academic language, which includes "book words," is the formal language of school. It's not the way anyone naturally talks. It includes abstract STEM concepts, nuanced multi-meaning words, literary devices, grammatical conventions, and more. Speak Agent, Inc. is dedicated to closing the academic achievement gap for ELLs and students with low socioeconomic status by promoting academic language learning integrated within every content area. This is what the company was founded to solve.
But why did we make Speak Agent a customization platform for teachers, rather than an adaptive learning program? After all, "adaptive learning" promises to automatically personalize content and scaffolding to exactly the right level for each student. Who wouldn't be for that?
My co-founder Ben Grimley and my colleague Dan LaFountain are both former ELL teachers. Speak Agent was largely born out of their frustration with the lack of effective tools for teaching academic language. Before I designed our first prototype, they'd already interviewed 50 fellow second-language teachers. All 50 teachers loved our ideas for engaging students with game-based learning and videos, but they also all shared one critical need:
They needed their students to grasp the key academic concepts for the current lesson and not be off on another track with computer-based learning. They needed a bespoke solution that fit into their curriculum and lesson plans and didn't require them to adjust their scope and sequence to fit a software product. And they needed a tool to augment their teaching, not to replace them.
As a software developer who regularly mentors other developers through DC PyLadies, I really value the human element in teaching. I wanted Speak Agent to keep educators in control, not software. After all, teachers know their students! Adaptive software only knows students based on data fed to it, which can lead to some truly ridiculous results. Adaptive software also cuts teachers out of the student-computer interaction. We saw the need for technology that is educator-managed and student-driven, rather than governed by unaccountable, uncaring software algorithms.
As my colleagues point out, academic language learning requires the three C's of communication, context, and collaboration. A program limited to student-computer interaction can miss some or all of those elements. Half of our product development team are former ELLs, so they provide a great balance to the half of the company that are former teachers. On this point, both student and teacher sides were unanimous: We need those three C's.
Then we added a fourth C for customization of the curriculum.
Speak Agent is not an adaptive learning product. It's a customization platform uniquely suited to academic language learning. We take your curriculum and build a tailored interactive experience for your students. You simply provide us with whatever you have—it could be lists of key academic concepts or unit vocabulary, selected reading materials, worksheets such as cloze activities, pacing guides, or any number of other documents—and then we piece together a customized program ready to start using within weeks. We can do this for any grade level and any subject. How can we possibly do this? Well, that is the technology I've spent the last three years of my life building! It's our secret sauce.
Katie is CTO and Co-Founder of Speak Agent, Inc. A leading expert in the Python programming language, Katie designed the Speak Agent architecture and wrote the very first line of code for the current platform. She is an accomplished technical author and keynote speaker worldwide, as well as a leader of the Washington DC chapter of PyLadies, an organization dedicated to training women in Python.