How Highly Active K-12 Educators Are Using AI Tools Like MagicSchool

By JP Martinez Claeys, Ana T. Ribeiro, Chris Agnew, and Susanna Loeb

Early findings from the top 5% active users

Artificial intelligence is becoming increasingly embedded in teachers' daily work. Yet we know little about how educators are using these tools and whether usage differs across teaching contexts.

This post is the first in a series examining usage patterns among highly active users of MagicSchool, one of the largest K-12 AI platforms. This first post examines how teachers are using AI tools and how these patterns differ across elementary, middle, and high school grade levels.

Using platform activity data from approximately 87,000 U.S. educators in the top 5% of MagicSchool users between May 2023 and November 2025, we explore which grade levels and subject areas these highly engaged users teach and how they interact with AI tools. Narrowing the analysis to highly active users allows us to focus on educators who have integrated these tools into their regular practice.

The data

We partnered with MagicSchool, an AI-powered platform that provides educators with more than  80+ tools for a wide range of teaching tasks, including instructional material creation, assessment generation, writing feedback, student and family communication, and administrative assistance¹. While many tools are designed for specific educational workflows, educators also use MagicSchool's multi-purpose AI assistant chatbot, Raina, for open-ended tasks.

This analysis focuses on U.S. self-identified educators in the top 5% of MagicSchool usage, defined by the number of distinct days they visited the platform between May 2023 and November 2025. This sample of 87,000 teachers who used MagicSchool on at least 52 distinct days includes information about the grade levels and subject areas teachers reported teaching on their platform profile, as well as records of their interactions with platform tools. Each interaction record includes the tool used (e.g., Lesson Plan, Writing Feedback), the associated thread (a separate conversation or tab used to generate content), and, depending on the tool, other fields that call for specific information tailored to the tool’s objective (e.g., grade level). In addition to user self-reported and interaction data, MagicSchool also provided a list of education and professional use categories associated with each tool – e.g., Multiple Choice Quiz / Assessment is classified as Feedback & Assessment, while Email Responder is classified as both Admin and Communication.

Who Are MagicSchool's Most Active Users?

Elementary and high school educators make up the largest share of highly active MagicSchool users, with English Language Arts representing the most common subject area.

Grade-level composition

Figure 1 shows the share of U.S. MagicSchool highly active teachers by grade level. Elementary and high school teachers are the largest groups of highly active MagicSchool users, making up 32.1% and 31.6% of the cohort, respectively. Elementary teachers make up the largest share of teachers in the U.S. overall (about 51%), yet are underrepresented among MagicSchool's most active users. In contrast, middle school teachers (18.2%), high school teachers, and especially "multi-level" teachers (18.2%) who report teaching across multiple grade levels, are overrepresented relative to their share of the teaching workforce.^{2 3}

Subject composition

Subject heatmaps

Figure 2 summarizes the subject areas reported by highly active MagicSchool users across middle school, high school, and multi-level teaching assignments. Elementary teachers are excluded because two-thirds identify as teaching "All Core Subjects," reflecting the generalist nature of elementary instruction rather than a distinct subject specialization. 

Across all three groups, English Language Arts (ELA) is the most commonly reported subject. Beyond this shared pattern, subject specialization varies by grade level. Middle and high school teachers are distributed across several core disciplines, including Social Studies, Science/STEM, and Math, whereas multi-level educators are much more likely to report Special Education alongside one or more additional subject areas.

Which AI Tools Are Highly Active Teachers Using Most?

While the top two tools used by teachers are highly flexible, such as the multi-purpose AI assistant Raina and the Text Rewriter, a tool for adapting text readability and length, AI tools designed to be task-specific and support classroom activities represent more than half of the 10 most used MagicSchool tools.

Top tools

Note: Tools shown in bold require a teacher to specify a grade level to generate content. Underneath each tool, a comma-separated list of tool categories is included.

Figure 3 shows, for the top 10 most-used tools, what percentage of all threads for power users belong to each tool. From the plot, we note that the top 10 tools together account for more than 60% of interactions from these users. Raina and the Text Rewriter are the two most-used tools by a large margin, with usage double or more than that associated with the third-place tool, Writing Feedback. These tools, particularly the multi-purpose AI assistant, differ from many others on the platform because they can support a wide range of teaching tasks rather than a single workflow. The higher usage of these tools is consistent with findings from our earlier work showing that educators often gravitate towards flexible, multi-purpose AI assistants.

For the remaining tools in the top 10 list, we observe three dominant types of tools: Feedback and Assessment, Admin & Communication, and Instructional Materials. The Feedback and Assessment category is associated with the two tools ranking 3rd and 4th in usage, standing out as the most-used category of tool after multi-purpose tools. Combined with three tools in the Instructional Materials category and the Lesson Planning tool, we find that AI tools designed to support classroom activities account for more than half of the 10 MagicSchool tools most used by power users.

Do Teachers’ AI Needs Change with Grade Level?

Educator AI usage shifts across grade levels. While Raina maintains consistently high usage across all grades, we observe differences in the types of task-specific AI tools that elementary, middle, and high school educators interact with the most.

Top tools by grade level

Note: Tools shown in bold require a teacher to specify a grade level to generate content. Underneath each tool, a comma-separated list of tool categories is included.

Figure 4 shows how the usage of each of the top 10 tools is distributed within each grade level. For each grade, we look at all threads and calculate what percentage of those threads are attributed to each of the 10 most-used tools (ranked by total threads overall).⁴

When we look closely at how the top tools are distributed in the previous figures, a distinct pattern begins to form. The data shows that teachers across grade levels have certain similarities as well as clear differences in the tools they use:

• Across the entire spectrum: Use of the multi-purpose AI assistant, Raina, is remarkably similar across grade-level groups, representing 17-19% of teachers’ interactions with tools on MagicSchool, suggesting that teachers at all levels value a flexible tool for open-ended tasks. 
• Where it starts to differ:
  • Elementary School teachers are the only group whose use of Text Rewriter surpasses the multi-purpose AI assistant, Raina, while their use of Writing Feedback is relatively low. 
  • Middle and High School teachers use Writing Feedback at nearly the same rate as Text Rewriter (around 10% of threads each), showing a clear shift toward task-specific grading and assessment as students get older. 
• Differences emerging from task-specific tools:
  • Although differences in use across grade-level groups become less pronounced for each individual tool as we move down the ranking, we can observe that teachers from different grade levels appear to be more drawn to task-specific tools from different categories. For example, tools used for admin and communication purposes, such as drafting emails, represent a larger share of Elementary School teachers’ interactions compared to other groups, while the opposite is true for tools in the Feedback and Assessment and Instructional Materials categories.

What Broader Patterns Characterize AI Use by teachers of different grade levels?

Educator AI usage reflects distinct workflow demands across grade bands. Elementary school educators concentrate tool usage on administrative work, communication, and student support, whereas middle and high school educators primarily use AI for subject-specific content generation, assessment, and grading. 

Note: Some tools are associated with more than one category (e.g., Text Dependent Questions is considered part of both Feedback & Assessment and Instructional Materials). Since the tool is considered to serve each assigned use case, a thread from a multi-category tool is considered in each category associated with the tool. Multi-category tools account for 7% of all interactions in the dataset. As a result, category-level averages may overlap and should not be summed to represent total platform usage. Task Support and Learning Support categories are absorbed with the Other category, as the number of threads in this category represents less than 0.1% of all threads in each grade level.

To understand whether these grade-level differences in use of task-specific AI tools hold true beyond the top 10 tools, we expand our analysis to all tools used by teachers in our sample and group them into categories of use that represent educational or professional domains (Instructional Materials, Communication, Feedback & Assessment, etc) that MagicSchool assigns to each tool.⁵ These domains allow us to identify the main purposes of why teachers use the platform. 

First, while the multi-purpose AI assistant chatbot was among the most-used tools overall, for every grade level other categories saw more intensive use than the Chatbots category. This suggests that teachers are choosing tools designed around their specific needs over the multi-purpose chatbot (Raina represents over 97% of all tool threads in the Chatbots category), even though the chatbot can be instructed to perform the same tasks, pointing to the value of tools built to fit a particular use case rather than a one-size-fits-all assistant.

Second, when inspecting what is differentiating each grade level, we find that Elementary school teachers use AI tools for admin and communication purposes more than any other group, as well as tools in the student support category (driven mostly by the use of the Text Rewriter tool). Middle and high school teachers, on the other hand, use tools that support academic content generation more, such as generating feedback to student work, creating quizzes and worksheets, writing instructional materials, and building lesson plans. Multi-grade teachers resemble elementary school teachers in their consistent use of Student Support and Admin and Communication tools, but differ in that they also make heavy use of academic content tools, where they are the second-largest users of the Instructional Materials category and the largest users of the Planning category.

The different patterns in the use of specialized AI tools from teachers in each grade-level group suggest an association between student age and academic development with the types of tasks their teachers engage with and may seek AI for support. For example, creating written exams or drafting feedback on assignments is not very useful if your students are still learning to read and write, but using AI to draft and customize messages to many families and caregivers about practicing reading skills at home might help improve literacy outcomes through family engagement. As students grow older and become more independent, direct communication between teachers, families and caregivers becomes less frequent, but instructional content becomes harder to learn and assessments more frequent, so teachers may turn to AI to make lessons more interesting and provide more information and direction to students when they make mistakes in an assessment.

Beyond this gradual shift from using AI for communication-intensive tasks to tasks focused on classroom activities, teachers of older students seem to be using AI to adapt or create materials to match the academic level of their students. We observe this pattern in Figure 6, where we separate AI tools that require grade information to be used, AI tools that don’t need grade information, and the multi-purpose chatbot Raina.⁶

Tools that require grade information are mostly focused on generating content specifically for instructional materials, feedback and assessment, while tools that don’t require a grade level tend to support general tasks.⁷ Because chatbot AI assistants don’t require a grade level to be specified, given that they can support different types of use cases, we treat these interactions as a separate category.

A clear pattern emerges across grade levels. Teachers in middle school, high school, and multi-level contexts engage more frequently with grade-specified tools, while elementary teachers make greater use of non-grade-specified tools. This pattern reflects a greater need in upper grades for tools designed to tailor content to specific academic standards and grade levels.

Essentially, these findings reveal a dual footprint across the platform. While all educators use general AI assistance at similar rates, their choice of specialized tools differs across teachers from different grade levels, with communication and student support tools playing a larger role in the lower grades and academic-content tools becoming more prominent in the higher grades. l

What Are We Learning?

Several themes emerge across these findings:

• First, the multi-purpose AI assistant is the most-used tool overall and is used consistently across grade levels, reflecting its broad value. However, at every grade level, certain use-case categories see more concentrated use than the Chatbots category — suggesting that tools built for specific instructional needs provide more value to the teachers who use them than a one-size-fits-all assistant. These patterns show that teachers consistently combine flexible, multi-purpose AI tools with specialized tools designed for particular instructional workflows.
• Many of the most common uses of AI involve routine but time-intensive professional tasks, including communication, student support, and administrative work. These use cases represent a higher use for elementary school teachers.
• AI use appears to become increasingly instructional as grade level rises. Compared to elementary school teachers, middle and high school teachers rely more heavily on tools that generate content, assessments, and feedback tied to specific courses and subjects. 

Together, these findings suggest that AI adoption is shaped not only by access to technology, but also by the everyday demands of teaching different students and grade levels.

Looking Ahead

This analysis represents an early look at how highly engaged educators are using AI tools within one platform. While grade-level trends establish an important baseline, macro-patterns only reveal part of the story.

In our next post, we will dive deeper into the data to explore how specific subjects change how teachers interact with AI, focusing on the distinct usage patterns between English Language Arts (ELA) and non-ELA educators to better understand how subject-matter demands shape daily teacher workflows.

¹ In addition to the standard tools offered by MagicSchool, teachers can create custom tools by adding instructions to modify the behavior of existing tools. Customized tools represent a negligible share of all threads analyzed.

² Teachers were categorized into the following mutually exclusive groups, (i) Elementary School: Teachers who reported 5th grade or lower as their highest teaching grade; (ii) Middle School: Teachers who reported to only teach grades between 6th and 8th, plus teachers who reported teaching multiple grades within the 5th to 9th grade range to accommodate variations in middle school grade compositions; (iii) High school: Teachers who reported 9th grade or higher as their lowest teaching grade; and (iv) Multi-level: Teachers who reported teaching multiple grades across two or more grade levels as defined above.

³ In contrast, when we consider the full sample of US Magic School K-12 teachers, 43% are classified as Elementary School, 16.3% as Middle School, 26.6% as High School and 13.9% as Multi-grade, which is closer to the distribution of US K-12 teachers, but still differ significantly for Elementary School teachers and Multi-grade teacher.

⁴ The top 10 tools represent for each grade level, 65.3% of all threads for elementary school teachers, 65.5% for middle school teachers, 64.7% for high school teachers and 64.5% for multi-level teachers.

⁵ For the purpose of the analysis by tool category, the “Admin” and “Communication” categories were grouped together into “Admin and Communication”.

⁶ In terms of specification, tools vary in the types of inputs they accept and the level of structure they require.. For example, Raina only requires free-form text to be able to generate materials or answer questions, but it can take multiple types of documents and other files as part of the prompt. Most other tools have a set of fields to direct the user to provide information that will serve as context to produce output with better quality and alignment to the teacher’s needs. For example, the Writing Feedback tool has three required fields – a description of the assignment, student grade, and the student assignment to evaluate– and two optional fields – a description of the type of feedback or rubric (with the option to attach a rubric), and additional instructions for providing feedback.

⁷ Share of threads with a grade level specified, by tool category: Feedback & Assessment, 98%; Instructional Materials, 96%; Planning, 96%; Student Engagement, 60%; Student Support, 31%; Learning Supports, 29%; Task Supports, 25%; Communication, 18%; Admin, 17%. Notably, in the Student Support category, the Text Rewriter, a tool highly used by elementary school teachers, doesn't require users to specify a grade.