Imagine a U8 soccer match where a swarm of young players cluster around the ball, each determined to kick it in any direction. It’s chaotic, enthusiastic, and largely unproductive until a coach guides each player into position, turning disarray into effective teamwork. Much like in these soccer games, the world of AI in education often has eager tech companies rushing enthusiastically towards innovation, but not always with a productive output. With proper guidance, EdTech companies can avoid creating solutions that, while technologically impressive, miss the practical needs of their intended users. These solutions are where educators come in, serving as the essential ‘coaches’ who provide the crucial feedback necessary to tailor AI tools that are advanced, appropriate, and useful for students, teachers, and families.
In this article, we’ll explore how Saddle River Day School (SRDS) embarked on developing and integrating an AI platform explicitly tailored to enhance educational outcomes. In the early months, the platform supported teachers in curriculum development for classroom use or integration into the school learning management system (LMS). By mid-school year, students could start testing the online lessons and assignments that were personalized or consistent for each student. Starting next fall, we plan to integrate parent access to the platform. By collaborating closely with teachers and students throughout the process, the project ensured the technology was innovative and aligned with educational needs, significantly impacting our teaching and learning environment while maintaining our strong sense of community.
The journey of integrating an AI platform at SRDS began with a pivotal trip to India in July, where we met with the development team that would bring our vision to life. These initial meetings were not merely technical but the foundational steps in crafting a tool that would resonate with our educators and enhance our students’ learning experiences. We outlined a robust plan that was ambitious yet grounded in realistic educational needs. It is important to note that our partner in the journey, the development company, was equally eager to build a tool with a significant academic impact.
Upon our return from India, we introduced the concept at the annual retreat for Department Heads in late July. The excitement was palpable, and the potential impact of AI on our educational practices sparked vibrant discussions. Our Division Heads identified potential Lower, Middle, and Upper school pilot teachers. Originally aiming for three teachers per division, the interest exceeded our expectations, with more than 20 teachers eager to participate. This enthusiasm was underscored by the community’s regular use of innovative educational tools.
In August, our faculty meetings took on a new dimension, including training sessions and an introduction to AI tools and platforms for interested faculty. The CEO of the development company (later named ProSchool) dedicated several weeks to working directly with our teachers, which was an invaluable engagement. During these sessions, teachers expressed their needs and desires, compiling a comprehensive list reflecting their aspirations and the practicalities of their daily teaching environments. This feedback became the cornerstone of the development efforts. This proactive approach in the early stages set a solid groundwork for subsequent phases of the project. As the school year commenced, we were prepared and energized to refine and expand the platform as the tools became available, ensuring it would genuinely support and transform our educational practice.
The fall semester was a period of intense development and collaboration for the SRDS.AI platform. Our teachers became pioneers, engaging in numerous Zoom calls as brainstorming sessions and practical workshops. These virtual meetings were crucial, as they allowed teachers to communicate their evolving needs and experiences with the platform, shaping its development in real time. In short, the more our teachers knew and saw, the more new ideas arose and became feedback.
As the developers assimilated the feedback, they gained a deeper understanding of the methods and requirements of the teachers. Unlike typical software development, this process was unique—there was no pre-existing model to emulate. All participants recognized it was impossible to rely on their own experience as students because none had ever been in classes using AI. Our platform was breaking new ground, necessitating reverse training where educators taught developers about the nuanced demands of an educationally sound environment. This collaboration was pivotal, illustrating a dynamic exchange of expertise where each Zoom call added a layer of refinement.
Throughout this iterative process, the developers worked diligently to enhance the platform’s capabilities, ensuring it aligned well with the pedagogical goals of our faculty. Hundreds of Zoom calls and meetings led to thousands of hours of coding and user-interface testing. They focused on optimizing the AI platform to complement and enhance the rich, curriculum-based work our teachers had developed over the years. The platform was not meant to replace traditional teaching methods but to augment them, adding value through personalized learning paths, data-driven insights, and scalable educational resources.
The synergy between our teachers and the development team was evident in the gradual transformation of the platform. Teachers started noticing that the AI could handle more complex educational tasks, such as adapting content to individual student needs and providing real-time feedback, which had previously demanded a significant portion of their time. This phase enhanced the platform’s functionality and solidified a shared vision of what AI in education could achieve. By the end of the fall semester, the platform was not just a tool but a testament to the power of collaborative innovation, developing a new standard for integrating technology in educational settings. The public launch at our annual State of the School address was exciting as our Head of School announced the many new advancements at the school.
The initial integration of our platform into the curriculum at SRDS closely mirrored other AI educational tools, providing a familiar ground for teachers to begin crafting and rethinking their lesson plans. However, as the platform evolved throughout the fall and into the spring, its capabilities expanded significantly, marking a transformative phase in curriculum integration.
Teachers began to utilize AI more dynamically by collaborating with it to develop assignments directly linked to their lessons. This process included the creation of detailed rubrics designed to assess these new AI-generated assignments, further enhancing the educational process. A particularly innovative feature was teachers’ training of AI assistants. These AI assistants guide students but do not answer their questions, significantly reducing the need for teacher intervention outside classroom hours.
The impact of these developments was profound. Some teachers reported up to an 80 percent reduction in preparation time and a noticeable decrease in after-hours student emails seeking help (which we interpreted as a potential sign of lower student anxiety related to the work, among other things). This freed up valuable time, which teachers then invested in direct student interaction, reviewing the content of the student-AI interactions, and refining and improving their lessons. Despite the high quality of AI-generated materials, our educators continuously sought to elevate the content further, ensuring it met our high educational standards.
During this period, we also introduced the “personalization” feature, a critical enhancement that allowed teachers to activate AI-driven customization for each lesson and assignment. This tool adjusted the content to fit each student’s learning preferences, abilities, and needs, creating personalized problem sets and learning paths. This level of personalization not only catered to diverse student profiles but also significantly enriched the learning experience, making education more engaging and effective.
Gradually integrating these sophisticated AI capabilities into our curriculum enhanced educational delivery. It set a new benchmark in personalized learning, demonstrating the platform’s growing potential to transform traditional educational models. The feedback from our faculty was instrumental in evolving the AI platform, directly influencing the addition of new tools and enhancements. A significant upgrade came when developers introduced a feature that allowed teachers to upload their previously created educational materials in PDF format. This functionality enabled the AI to learn from a vast repository of existing content, which it used to refine and expand the lesson plans further.
Around the same time, we observed a notable expansion in the platform’s language capabilities. Initially designed to support major global languages, the platform soon began accommodating lessons in different languages, from Romanian to Farsi or Hausa. This development was a technical achievement and a broadening of our educational outreach, allowing us to cater to a diverse student body with varying linguistic needs and a range of world language offerings.
Despite occasional setbacks—such as AI malfunctions during live demonstrations—our faculty’s response was remarkable. They embraced these challenges with resilience and a forward-thinking attitude, always eager to explore further improvements. Their feedback was not just reactive but also visionary. It was often punctuated with inquiries about future possibilities, reflecting a continuous drive to push the boundaries of what our AI platform could achieve.
This dynamic feedback loop, characterized by robust teacher engagement and responsive development, ensured that each update addressed immediate issues and opened new avenues for educational innovation. As our AI platform evolved, it distinguished itself significantly from typical AI educational tools. Initially perceived by outsiders as just another lesson creator or AI tutor, the platform’s true potential gradually became apparent to our faculty and the wider educational community. Discussions between our teachers and their peers sparked interest and curiosity, highlighting a growing awareness of the platform’s unique capabilities.
The AI’s ability to develop advanced educational content was a revelation. It successfully created college-level courses, including a graduate school finance course, and specialized content, such as a training program for registered nurses. In an impactful instance, the ProSchool platform developed a course in Yoruba for a Nigerian NGO designed to explain the risks of dangerous sea crossings in the Mediterranean in small boats. These examples illustrated the platform’s versatility and potential to make significant educational contributions beyond traditional school settings.
We have shared the knowledge we are acquiring with many schools and NGOs at various regional, national, and international conferences. The NGOs are particularly interested in how AI might transition rural education worldwide, particularly with the expansion of low-earth-orbit satellite systems that will not require cell towers or fiber optic cables for internet access in the most remote regions. Building on our success, we are preparing to launch an ambitious pilot program to support refugees. This summer, the ProSchool platform will offer educational continuity in environments where traditional education is disrupted or inaccessible. Whether due to constant relocation, language barriers, or the lack of stable educational infrastructure, the AI platform will provide tailored educational content to meet the diverse needs of refugee students.
Expanding the platform’s scope demonstrates technological innovation and a commitment to leveraging this technology for social good. The ability to adapt and create relevant content across a spectrum of subjects and needs underscores the profound impact our AI initiative is having on the global educational landscape. Additionally, we have noticed that AI can target lesson creation to reflect government regulations that teachers in other regions and school systems must meet.
Throughout the development and integration of our AI platform, one of the most significant lessons learned was the importance of continuously expanding and adapting technological tools to meet the specific needs of educators and students. The introduction of speech-to-text and text-to-speech functionalities exemplifies this evolution. Teachers found it increasingly effective to communicate directly with the AI bot via voice, enhancing their ability to build and modify content efficiently. Similarly, students benefited from the text-to-speech feature, which allowed them to listen to lessons in sync with reading, aiding comprehension and retention, especially for learners with auditory preferences.
Our developers tailored these audio tools with different voices and speeds to suit various age groups, enhancing accessibility across our diverse student body. For our math teachers, integrating LaTeX code capability ensured that complex equations were accurately represented, while lower-school teachers could present traditional fractions effectively. Another innovative development was optical character recognition (OCR) technology to read students’ handwritten submissions, which proved invaluable for assessing written assignments more efficiently.
As we approached exam seasons, the platform was instrumental for AP teachers to generate extensive practice materials, helping students prepare effectively. These enhancements not only improved educational outcomes but also demonstrated the potential for AI to reduce educator workload and enhance student engagement significantly.
For educators seeking to embark on similar AI initiatives, our advice is to prioritize flexibility in development and maintain an open dialogue with all stakeholders. Such collaboration is crucial in ensuring innovative technology solutions are tailored to educational needs. AI has evolved steadily over the last few years, and an iterative design process for any tool will be essential to embrace the available changes.
Throughout our journey of integrating a pioneering AI platform at Saddle River Day School, we’ve seen transformative changes in how we teach and our students learn. The visionary leadership of the Head of School, supported by the Board of Trustees, has been instrumental in fostering this forward-thinking adventure. Advanced tools such as speech-to-text, text-to-speech, LaTeX integration, and OCR for reading handwritten work have enhanced educational experiences, streamlined teacher workflows, and improved student engagement. As we look to the future, the possibilities for further development are vast. This summer, our faculty, led by our newly appointed AI strategist, will continue to refine and expand the platform’s capabilities. With ongoing visionary support and a commitment to innovation, we will further explore the transformative potential of AI in education.
About the author
Dr. Kevin Merges is Saddle River Day School’s (SRDS) Chief Global Impact Officer. At the beginning of his career, he taught secondary school mathematics and later at Rutgers University. He has presented at conferences around the United States and over a dozen countries worldwide on communication through online environments, international education, artificial intelligence, and virtual reality to communicate education to constituents. His current work is expanding global initiatives and revenue streams for SRDS.
