For tech-forward education companies on the forefront of building artificial intelligence tools, the advancement of agentic AI could sound like the path to a more autonomous future for teaching and learning.
Agentic AI aims to be a semi- or fully autonomous form of technology that is able to perceive, reason, and act on its own with limited human prompting. Within the space of K-12, agentic AI could look like launching customized lessons when concepts aren’t fully grasped, nudging students at risk of falling behind, or generating provisional feedback on essays.
But while the use cases may sound intriguing, one of the biggest challenges for ed-tech companies trying to incorporate agentic AI into its offerings may come very early in the process — before schools even decide whether to buy the technology.
Don’t Miss the Summit Everyone Will Be Talking About
Join the education industry leaders shaping what’s next. Be there LIVE in Nashville, November 11–13, for exclusive insights from district decision-makers, the latest shifts in the K-12 market, and the connections that can drive your organization’s growth.
New data from EdWeek Market Brief found that many educators are still trying to understand what agentic AI actually is, and how it would fit into their work.
In an online, nationally representative survey conducted by the EdWeek Research Center from March to May, 753 educators — including 116 district leaders, 41 school leaders, and 596 teachers — were asked about their experience with the use of agentic AI for work purposes.
According to the data, a quarter of respondents say they have never heard of agentic AI. About 18% have heard of it, but they aren’t sure whether they had used it for work.
About a quarter of respondents say their experience with the technology is neutral, while 14% say it is negative, and 18% say it is positive.
The overall unfamiliarity with agentic AI and lack of awareness of whether respondents have been using it could present a challenge for vendors seeking to bring agentic AI products into K-12 schools.
With agentic AI products just now being developed and emerging, there is also little visibility into how they would work alongside current products or ed-tech stacks.
What Makes AI ‘Agentic’?
Unlike generative AI, which typically responds to a specific prompt, agentic AI is designed to take a series of actions toward a goal with less human direction.
For example, a generative AI tool might create a lesson plan after a teacher asks it to. An agentic AI system could potentially identify that a group of students is struggling with a concept, select or create an appropriate lesson, assign it to those students, monitor their progress, and then adjust the next steps based on their performance — all with limited human prompting.
That added autonomy is what distinguishes agentic AI from many of the AI tools educators are familiar with today. It also raises new questions for schools about how much decisionmaking they are comfortable delegating to a technology system.
Lindsey Clements is among the educators who are familiar with AI, but have never heard of the term “agentic AI.”
As a 1st grade teacher in the Ritenour School District in Missouri, Clements uses relatively little AI in her classroom, although the technology is already present in some of the tools and curriculum-development processes used by her district.
She doesn’t generally seek out AI tools, but she considers herself comfortable with technology as a whole, especially since she works a part-time job at Apple.
“I think [agentic AI] can be really helpful,” she said. “But as somebody who learned how to teach without AI, it’s just not something I think about or something I incorporate into my day-to-day.”
That distinction could be important for vendors. Teachers may not necessarily be resistant to technology itself; they may simply not yet see agentic AI as something they need.
Scott Kemple, head of school at Kalamazoo Country Day School in Michigan, said he sees significant potential in agentic AI, but believes educators are still in an early stage of understanding the technology.
At Kemple’s school, an independent, non-profit private school with a little more than 100 students, teachers have been provided AI training since the early days of ChatGPT, but adoption has been uneven, he said. He’s seen some teachers use AI to develop curriculum and manage data, while others have stayed away from it.
But the difference becomes even more pronounced when moving from generative AI to agentic AI, he said.
“Agentic AI is so new,” Kemple said. “There’s distrust in this new form of AI that is very independent, that is its own problem-solver, so [teachers] would need to know what are the safeguards and is my data safe?”
For vendors pitching agentic AI, a distinction will have to be made between an AI system that generates a response to a prompt and one that can independently carry out multistep tasks, Kemple said, as well as what that means for the classroom.
Sentiment Leans Negative
EdWeek Market Brief’s survey also asked respondents, if a vendor announced the development or use of agentic AI, whether the likelihood that they’d recommend using that product in their district or school would increase, decrease, or have no impact.
About 43% of survey participants say their likelihood to recommend that product would decrease. Twenty-two percent say it would increase, and 35% say it would have no impact.
The unfamiliarity aspect could play a role as to why people aren’t as comfortable using or recommending an agentic AI tool, Kemple said.
“When you talk about an AI that has that much independence, that much problem-solving ability, that much freedom, I can see where that would be scary for a lot of people,” he said.
Clements said she’s particularly uncomfortable with the idea of an AI system making decisions based on limited information.
“The idea that something is going to make a decision based off computer input is just scary,” she said.
District vs. Teacher Perspective
When broken down by role, however, the survey results revealed a significant difference in how district leaders and teachers view agentic AI.
Among district leaders, 39% say an announcement that a vendor was developing or using agentic AI would increase their likelihood of recommending the product. Just 17% say it would decrease their likelihood.
Teachers were much more cautious. Only 19% say the announcement would increase their likelihood of recommending the product, compared with 48% who say it would decrease their likelihood.
The divide may partly reflect the different ways the two groups encounter new technologies.
District leaders may be thinking about the potential of AI to improve efficiency, support teachers, or change how work gets done across a system. But teachers have to figure out how a new tool fits into an already crowded school day, Clements said.
From an administrative perspective, AI can make it possible to create lessons more quickly, she said. But from a teacher’s perspective, that also means still having to review the material, determine whether it’s accurate, and making sure it fits their instructional goals.
“I would need to see that people are still behind [the decisions],” when it comes to what would make her feel comfortable using agentic AI with her elementary-level students, Clements said, “[so] that the decisions made and the pieces created are from people and that it aligns with my knowledge, and that the company stands for the same thing that my district stands for, which is making sure it’s what’s best for kids.”