AI Agents and the Future of Digital Labor in K12
2025 has been dubbed the “Year of the AI Agent” by tech giants like Marc Benioff and Sam Altman. What could this mean for K12 education this year and in the years to come?
Created using ChatGPT
The launch of ChatGPT in November 2022 ignited public interest in generative AI. A little over two years later, we’ve seen the impact of large language models (LLMs) from companies such as OpenAI, Anthropic, Google, and Meta as they help millions of people with everyday “AI Assistant” tasks including writing emails, summarizing documents, and brainstorming ideas. We’ve also seen an explosion of application layer tools built on top of LLMs tailored to the needs of consumers and users in those industries. Examples include ChatGPT, Perplexity, character.ai, and Claude.
All the buzz is now about “AI Agents” as the next great frontier. These agents go beyond the current AI assistants by enabling AI to proactively work towards goals without constant human prompts. Assistants handle tasks that are prompted by users, while agents handle a set of tasks or workflows without the need for human direction. Capabilities can include strategizing, problem-solving, and executing tasks autonomously. Early examples include Anthropic’s “computer use” to operate online interfaces like a human; Salesforce’s AgentForce automating workflows in customer service, marketing, and sales for their customers; and Replit’s agent for software development.
AI Agents in K12 Education
We’ve seen an explosion of task-oriented AI assistant tools in education across a wide range of use cases. Some founders are pursuing the holy grail of student-facing edtech: an affordable personalized tutor for every student. Other tools assist teachers with assessments and feedback, teacher practice support, instructional materials, teacher professional learning, and more. The most adopted tools to date (like MagicSchool and ChatGPT) save teachers time and effort in daily and weekly tasks that are an integral part of the job of teaching. With hundreds of AI assistant tools on the market now, many of these first-generation tools are useful but are limited point solutions. With increased competition and rapid changes to AI technology, many of these tools will struggle to reach meaningful scale and will likely be replaced in time by more advanced tools.
As I write this, none of these assistant tools act as AI agents and autonomously handle workflows yet. I expect that to change and have a strong feeling we’ll see attempts at autonomous classroom agents in the near future. A widely accepted belief in traditional K12 schools and districts is that the teacher is a critical, indispensable resource in the teaching and learning process and that learning is a uniquely human and social activity (I agree with this belief). With that, it’s an open question as to if/how autonomous AI agents will be adopted in traditional schools and districts, especially those that would replace a teacher in the workflow. I certainly have my doubts. What’s more likely is that we’ll see classroom AI agents emerge on the edges of education in areas such as the homeschool and microschool markets and direct-to-consumer options where resources are more limited and the opportunity to reimagine education is more open.
AI in School/District Administration and Operations has Yet to Emerge
While we’re seeing AI innovation in classrooms, innovations in the back office functions of K12 schools and districts have been slower to emerge. There have been add-ons to existing administrative software tools such as PowerBuddy from PowerSchool, but we have yet to see AI-native tools emerge that address systems and processes in school administration and operations. Many school administrators and staff handle daily and weekly jobs in human resources, transportation, food service, facilities, procurement, and finance using basic off-the-shelf tools (e.g., spreadsheets, forms, or paper and pen) or rudimentary software from the 1990s or 2000s. With limited software budgets for some of these functions, the addressable market has been too small to attract much in the way of investment capital or talented technologists to build consumer-grade software.
That could change with AI agents that serve as digital labor. While software budgets account for just 3% of school district spending, labor budgets can be as much as 75–80% of the budget. If you have AI agents that handle workflows traditionally done by human labor, then you can imagine a much larger market opportunity. By way of illustration, if agents are able to save just 6% of the labor costs that districts incur, district budgets for software/AI agents would nearly triple from 3% to 9% of spend. That presents a large opportunity for AI founders.
Another aspect that makes the opportunity for AI agents in school/district back office functions intriguing is that school districts are under heavy budget pressures and many are short-staffed. Pressures on local, state, and federal revenues that fund school districts along with declining enrollments provide little hope that budgets will meaningfully increase in the years ahead. Staffing cuts are happening in K12 schools and districts and the prospects are dim for reversing that decline. Often, the remaining staff are overburdened by the number of tasks/jobs they have to complete. When I was an operator serving HR departments in K12 schools and districts, a common refrain from staff was that they had many more things on their to-do list than they had time to do during the day. AI agents could help ease that burden without replacing human labor; the humans would be free to handle the more important, higher-function tasks while leaving the AI to handle the repetitive drudgery inherent in most all jobs.
For example, some of these administrative tasks are highly repetitive and relatively low stakes if and when an error is made, meaning they could be tailor-made for an AI agent to step in and complete them. In this scenario, the AI agent could complete the workflow autonomously while a human oversees the agent and makes corrections as needed. Imagine a school employee overseeing a series of AI agents that complete various workflows in human resources. One agent automates job postings, pre-screens resumes, and schedules interviews with human supervision while another agent handles the onboarding process for new hires and a third agent prepares payroll and handles employee benefit inquiries all under the supervision of a human HR manager. School districts can accomplish much more with fewer dollars in the back office, which frees up more funding to go directly into classroom instruction or more skilled staff needs.
A simple Venn Diagram for thinking about opportunities for AI agents in K12 education
Challenges and Considerations
As with the public at large, school leaders struggle to accept AI as part of the workforce. There are concerns about AI displacing workers in schools. Successful initiatives will position AI as an augmentation for existing staff that helps them do their job more effectively and efficiently but does not replace staff. I firmly believe that AI is much more likely to change jobs than replace them entirely. Rightly, there are also concerns over student and employee data privacy in schools. Finally, there are legitimate concerns about bias in AI that need to be addressed. We must ensure that the AI is trained to eliminate bias, which requires transparency in the data used for its training. Those issues need to be addressed before widespread adoption of AI agents will happen in K12 schools.
Conclusion
Industries such as sales and marketing, coding, and customer service are poised to be transformed by AI agents. Whether AI agents come to K12 schools and districts depends on how we balance technological advancements with human workforce, data privacy, and data bias concerns. Certainly, there are many use cases where AI agents could help teachers and students advance teaching and learning and help school administrators and staff operate schools more efficiently and effectively. Schools that adopt responsible AI agents as they emerge could simultaneously address challenges they face including budget deficits and staffing shortages. The business opportunity for talented education technology founders could be immense, especially if they’re able to move beyond traditional software budgets and tap labor budgets with their “digital labor” AI agents. If you’re thinking about and/or building AI agents for use in K12 schools, I’d love to hear from you.
