The “Hidden” AI Tools That Are Driving Educational Gains
Tutoring may seem straightforward, but scaling it effectively is a complex and costly endeavor. Estimates suggest that intensive, high-impact tutoring can range from $1,500 to $2,500 per student. For a school with 250 students in need of such support, the cost could exceed $500,000 — a figure that is often beyond the budget of most schools, despite compelling evidence that tutoring is one of the most effective strategies for boosting student achievement.
However, tutoring only yields significant results when tutors are both knowledgeable in the subject matter and skilled in teaching methods. Assigning students to untrained or inexperienced tutors typically does not lead to meaningful learning progress. On the other hand, when tutors are well-trained and adept at engaging with students, the academic gains can be substantial.
This is where Saga Education plays a pivotal role. Well-regarded in educational circles, Saga’s high-impact tutoring model is backed by extensive research.
Recently, Saga introduced an AI-driven platform designed to enhance tutor effectiveness. Developed in collaboration with researchers from the University of Colorado, this platform is part of the Learning Engineering Virtual Institute, a collaborative initiative across multiple organizations aimed at significantly improving math outcomes in U.S. middle schools. My organization, The Learning Agency, is involved in administering this virtual institute, so I’m familiar with the project.
Research has shown that the way tutors communicate with students can significantly influence their learning outcomes. Certain “talk moves” employed by tutors can maximize their impact.
Get Muhammad Subhan’s stories in your inbox
Join Medium for free to get updates from this writer.
One such technique is “pressing for accuracy,” where the tutor asks the student to explain the concept or idea they just discussed, ensuring that the student truly understands it. Another effective method is “pressing for reasoning,” which involves prompting students to articulate the reasoning behind their answers.
For instance, a tutor might ask the student to provide an example or explain their thought process by asking, “Why did you choose this method to solve the problem?”
Tutors who consistently use these techniques tend to achieve better results than those who do not. This is where AI steps in to help tutors improve.
After a tutoring session on the Saga platform, the AI analyzes the conversation between the tutor and the student. It evaluates who spoke more, the quality of the dialogue based on various talk moves, and the types of questions asked and how they were answered. The platform then generates visual reports, including timelines showing when certain talk moves occurred, charts displaying the frequency of these moves, and other data points, such as who dominated the conversation — the student or the tutor — and whether the tutor allowed one-word answers.
These AI-generated insights provide tutors with specific, actionable feedback for improvement. Typically, it would take a coach or teacher about 90 minutes to watch and annotate a 30-minute session to provide the same level of detailed feedback. The Saga platform reduces this to just 30 minutes.
This efficiency enables one coach or teacher to effectively oversee multiple tutors, increasing the reach of high-quality, intensive tutoring while managing costs.
The true promise of AI in education lies in empowering educators with the data, analysis, and tools they need to excel. It’s a technology that most parents may never notice, but their children could be benefiting immensely from it.
So, the next time you encounter skepticism or doubt about AI’s role in education, remember that AI is much more than just ChatGPT and automated worksheets. The “hidden” tools like Saga’s platform represent the real future of AI in schools.
