Education in the AI era - Part #1 : AI in K12
K12 classrooms are at an exciting juncture. AI-enabled tools, designed specifically for K12 education, are increasingly making their way into schools, offering potential solutions for easing teacher workloads, boosting student engagement, and improving administrative efficiency. AI has been present in some schools for years (think recommendation systems and reminder prompts…), but its role is becoming more prominent as K12 education systems grapple with issues like teacher burnout, widening learning gaps, and increasing administrative complexity. However, how much impact is AI really having in K12 classrooms? Which applications are gaining traction, and what barriers still remain?
In this blog, we’ll explore these questions, offer potential winning plays that can help startups find success, and look ahead to the role AI may play in reshaping K12 education.
The blog is split into the following sections:
Let's dive in!
Is AI really making headway in K12 systems?
While AI is finding its way into K12 , its impact remains varied and its potential largely untapped.
A 2024 HolonIQ report indicates a modest increase in AI adoption, with 30% of surveyed education institutions successfully deploying AI solutions, up from 25% the previous year. However, the OECD's Digital Education Outlook 2023 reveals that this adoption is skewed towards AI-enabled applications such as automated grading systems, attendance tracking, and resource management due to their immediate, measurable benefits. More advanced AI applications, like personalised learning platforms and predictive analytics, remain underutilised, with challenges such as inadequate infrastructure, insufficient teacher training, and concerns over data privacy slowing down broader integration.
This variance in adoption is reflected in perceptions across different stakeholders. While 41% of school administrators support AI integration, only 21% of teachers currently share that enthusiasm, as per Frontline Education's 2023 Brief.
This highlights the mixed perceptions of AI's utility, particularly in day-to-day classroom activities. In schools where these advanced tools have been piloted, initial studies suggest improvements in operational efficiency, but the jury is still out on their long-term impact on student engagement, learning outcomes and whether perceived reductions in teacher agency in areas of their role affected by AI are positive or negative.
Overall, K12 schools remain cautious, with the OECD report highlighting the need for more evidence of AI's tangible educational benefits before committing to widespread implementation. Despite these challenges, startups continue to innovate and build in the K12 space. Encouragingly, there is also a growing emphasis on AI literacy, with some European countries developing AI curricula for high schools to prepare students for an AI-driven future. Government initiatives are also playing a crucial role, with some, including the French government, launching ambitious strategies to integrate AI across their education systems. These efforts suggest that while AI's current impact in K12 education may be limited, the groundwork is being laid for more significant integration in the future.
Betting big: where are startups placing their AI chips?
Despite the uneven pace of AI adoption in K12 schools, startups are bullish about AI’s potential to tackle some of education’s biggest challenges. While full adoption and integration may be slow and uncertain, the AI tools being built offer a window into what the future of education could look like.
The market map below reveals where AI startups appear to be focusing their energy in the K12 space.
--
Along the y-axis, we consider the stakeholder groups (learners, educators, and administrators) as well as the pain points, goals and AI use cases for solutions in the appropriate spaces.
Along the x-axis, we focus on the maturity of VC-backed startups in the space, taking the amount raised as a proxy for said maturity - we have 3 buckets which are $0-3M (and N/A which means unclear funding amount), $3-10m and $10m+.
Some companies address more than one of our 'use case' categories - in these cases, we have placed the company in the area that is most heavily featured on their website.
We will revisit this map when we discuss opportunities moving forward.
--
Startups in the K12 space have widely embraced content creation, with companies like Magic School (US), Curipod (NO), and Teachino (AT) making it easier than ever for teachers to generate lessons, quizzes, and interactive presentations. These companies are already claiming significant adoption—Magic School, for instance, reports nearly 1.5 million educators onboarded in just nine months, Curipod has reached approximately 5 million students, and Teachino is now used daily by thousands of teachers, schools, and educational institutions.
While these tools reduce the time spent on lesson planning and preparation, the market for solutions of this ilk is extremely competitive within K12 - given this market density, a focus purely on content creation may no longer be seen as a defensible differentiator for AI-first startups or a sufficiently broad vision for those companies seeking VC backing. This is not to suggest that startups focusing solely on content creation tools lack value or potential, but rather that their success may potentially be confined to local, regional, or niche markets, and that they might find it difficult to scale into larger, global enterprises given the competition.
One potential way to differentiate is by branching out into more advanced AI applications and features that facilitate personalised and adaptive learning paths or even interactive and immersive learning. Companies like EvidenceB (FR), Tassomai (UK) and Mangahigh (UK) are leaning into personalised and adaptive learning paths, offering tailored educational experiences that adapt in real-time to a student's progress, strengths, and areas for improvement. These solutions promise a level of personalisation that goes beyond content generation by actively responding to individual learning needs and tailoring content accordingly. Companies like Zzish (UK) and Brainspark Games (UK) are enhancing their content creation tools with interactive and immersive content, creating gamified learning experiences and adaptive flashcards that adjust based on a student’s real-time performance.
As AI’s role grows, solutions that blend content creation with adaptive learning and immersive experiences may find a more sustainable foothold.
The Bottom Line: Will K12 Schools Open Their Wallets for AI?
While AI in K12 education holds immense promise, the real question is whether K12 schools—often working with tight, pre-allocated budgets—are willing to dedicate funds to these solutions. With much of their spending locked into salaries, facilities, and core educational resources, any new investment must deliver tangible, measurable benefits that align with the immediate needs of schools. According to the OECD Digital Education Outlook 2023, many K12 schools are cautious about adopting AI technologies where the ROI is less well defined, less tangible or slow to materialise.
For example, tools enabling automated grading offer significant reductions in teacher workload– Sainaptic’s AI, for instance, is capable of marking qualitative exam responses in seconds with great accuracy. But while these tools and others can alleviate burnout and improve teacher retention, school budgets are limited and so tools are increasingly looking to exam bodies and consumers for revenue generation.
According to McKinsey's 2020 report, approximately 20-40% of teacher tasks- including preparation, administration, and grading- can be automated using existing AI tools, potentially freeing up 13 hours per week for teachers to focus on more meaningful student engagement. Despite this, teacher-to-classroom ratios tend to remain relatively fixed with school systems in the UK and Europe consistently spending at least 75% of their budgets on teaching and support staff.
So, while AI can automate certain tasks, the resulting operational efficiency may not be commensurate with the cost savings, meaning that these efficiency gains may not provide sufficient incentive for K12 schools to invest heavily.
Most educational institutions are by nature conservative, prioritising proven pedagogical impact before adopting new technologies. This conservatism is heightened by the fact that many districts already manage an overwhelming number of applications; for example, in the U.S, K12 school districts are on average using around 2,500 different tech applications. Establishing the efficacy of new AI applications is a slow process—"Right now, there is an ongoing RCT of Khanmigo in Canada, but we'll likely have results in a few years, not months…" noted Stéphan Vincent-Lancrin, Deputy Head of the Centre for Educational Research and Innovation (CERI) at OECD.
As a result, revolutionary adoption that fundamentally changes learning experiences remains limited. This challenge is likely pronounced for ambitious AI tools focused on personalised learning paths or interactive content, which promise significant student outcome enhancements but require upfront investments and have delayed returns.
For K12 institutions, the challenge is in balancing long-term gains with the immediate constraints of tight budgets, meaning that startups that can effectively demonstrate both immediate efficiency gains for administrators and educators, alongside measurable, longer-term improvements in student learning outcomes might be better positioned to secure a place in school budgets.
Unsurprisingly, AI products finding the most success seem to do so by integrating into the existing framework of schools rather than attempting to revolutionise education, according to Stéphan Vincent-Lancrin. In a conversation regarding this research, he noted that the AI tools selling well in K12 tend to replicate (or at face value, digitise) familiar practices and streamline traditional classroom methods. Rather than radically overhauling how education is delivered, these tools augment the established pen-and-pencil approach, making adoption easier for educators. This incremental improvement aligns with the cautious nature of K12 institutions, which often hesitate to implement drastic changes. This would align with the success Magic School (US) has had in reaching and being used by several million teachers in their lesson and activity planning within a relatively short period.
Additionally, AI startups face another hurdle: large, influential, recognised, trusted incumbents like Google for Education and Pearson dominate the K12 space with comprehensive, well-established platforms. These incumbents offer all-in-one solutions that integrate into existing school infrastructures, leaving little room for smaller startups to compete.
Winning Plays: How Are K12 AI EdTech Startups Making It Work?
Despite the hurdles to AI adoption in K12 schools, many startups are finding ways to make AI both feasible and valuable. We will preface by saying that there is no silver bullet or single "right way" to win in K12. Startups are finding success by using a variety of strategies tailored to the complex and varied needs of schools.
The opportunities below are based on our analysis of the market map, conversations with founders and experts in the ecosystem and our own perspectives.
Here are some of the key ways we see them finding success:
Using compliance as a hook: Startups like GoGuardian provide AI-driven tools for monitoring student online activity, helping schools meet GDPR and other safeguarding regulations. By addressing mandatory compliance needs, these solutions become indispensable for schools- they provide pain relief, not a vitamin.
Reframing AI products to fit non-tech budgets: Startups like TeachFX (US) (a Brighteye portfolio company) are able to position their AI tools as part of teacher training and development. By providing AI-driven insights into teaching, they allow schools to tap into professional development funds rather than constrained tech budgets. This approach to communications and sales could be applied to other areas touched by AI products.
Focusing on Immediate ROI: Zen Educate (UK) (a Brighteye portfolio company) leverages AI to streamline the process of finding and placing substitute teachers. By using AI to match schools with qualified substitutes more efficiently, Zen Educate provides immediate cost savings and operational efficiencies, which directly impact school budgets. This makes solutions like Zen Educate particularly appealing, as they deliver tangible benefits right away.
Finding Corporate Sponsorships and Alternative Funding: Some startups are tapping into corporate sponsorships and partnerships to fund their offerings, especially for schools with limited budgets. For instance, Memby (LT) (a Brighteye portfolio company), an AI-supported tutoring platform, partners with businesses to subsidise their costs, allowing schools to provide tutoring services at a reduced price or even for free. For corporations, this approach can drive sales by implementing models such as "buy one, give one" – where each product purchase helps fund an hour of tutoring for a student through the partnership. This not only generates revenue but also aligns the company with meaningful social impact, boosting brand loyalty and visibility. This doubles as an effective GTM approach in new markets.
Going Upstream by Targeting Companies Selling Educational Products to Schools: Startups can go upstream by partnering with companies like publishers that already sell to schools. Taskbase (CH), for example, integrates its AI tools into publishers' educational materials, leveraging their established distribution channels and relationships to scale more efficiently without selling directly to schools.
Adopting a B2B2C Model: Platforms like GoStudent (AT) introduce AI-powered tutoring through schools but pass the costs onto parents for premium features. This model allows schools to access innovative tools without tapping into limited budgets, while parents invest in enhanced learning for their children.
Targeting private and international schools: Private and international schools often have more budget flexibility and may be more open to adopting AI solutions. Startups offering innovative tools may find that these schools are more willing to experiment with new technologies, making them a promising starter market for AI products.
Strategic Market Entry via M&A: Another possible channel is entering a new market indirectly by leveraging existing brands or solutions already in the market. Instead of pushing a new brand directly into new markets, successful companies like Sdui (DE) (a Brighteye portfolio company) acquire local players that have an established presence. This enables a smoother transition and better reception. For smaller startups with limited resources, merging with established players or partnering with well-known local entities can be an efficient strategy to gain market entry and build credibility with schools.
Leveraging Political and Social Capital for Adoption: In Europe, particularly in Germany, securing buy-in from policymakers by aligning AI solutions with political priorities like equity and social integration can be a powerful strategy. When politicians and administrators see the societal benefits of these tools, they can help fast-track adoption in schools, using their influence to push through policy changes or provide funding that accelerates the startup's rollout.
The elephant in the room in this list is impact – given the enormous variation in how impact is defined and assessed from solution to solution, proving impact on learning or staff outcomes does not necessitate commercial success and scaling within K12 markets. This is not to say it won't in time!
AI’s Next Class: What does the Future Hold for AI in K12?
As AI continues to gain traction in K12 education, its future presents both exciting opportunities and unresolved challenges, particularly in addressing key barriers such as proving pedagogical value, securing willingness to pay, and impacting teacher-student ratios.
Here’s a look at what the future might hold:
AI tools evolve from assistance to co-teaching: Imagine AI not just as a tool for grading or lesson creation but as a co-teacher in the classroom. AI’s ability to provide real-time feedback to both teachers and students could transform how learning happens. Instead of waiting for test results, AI could give students instantaneous insights into their progress, suggesting new learning pathways based on their strengths and weaknesses. Startups focusing on AI tutors, real-time feedback systems, or AI-driven assessments will be well-positioned as schools look for tools that complement human teachers rather than replace them.
Personalisation at scale transforms the learning experience: In the near future, hyper-personalised learning might become the norm. AI will analyse not just academic performance but emotional responses, engagement levels, and even non-verbal cues. This data will then allow teachers to tailor learning experiences with unprecedented precision, ensuring that each student receives the support they need. Startups that can create adaptive learning environments capable of analysing multifaceted student data beyond academics will lead this charge. However, navigating privacy concerns and ensuring ethical use of student data will be critical to earning schools’ trust.
AI-powered schools reshape all facets of education: Another opportunity is the concept of entirely AI-powered schools, which would represent a new type of educational institution. These schools would leverage AI for everything from curriculum development and personalised learning paths to administrative management and student engagement. This model is particularly promising for special education, where tailored AI solutions can make a significant impact, aligning with the billions of dollars the U.S. government is investing in this area and that European governments may replicate in the coming years. By reimagining the entire school system around AI, these schools could address barriers related to proving pedagogical value and impacting teacher-student ratios.
AI-Powered Emotional Intelligence and well-being support becomes integral: A promising frontier is the integration of emotional intelligence into AI systems. Imagine AI tools that don’t just analyse what students know but how they feel. By gauging a student’s emotional state, AI could support mental health, boost motivation, and offer encouragement when students are struggling. Startups that integrate AI with emotional well-being tools—helping both teachers and students navigate the stresses of learning—could transform how schools view AI, from a learning tool to an emotional support system that forms the basis of AI-powered counselling. There is a rising wave of companies connecting students to mental health and wellbeing guidance, such as Cartwheel (US) - these companies likely pave the way for further AI-powered solutions in this space in the coming months and years.
Multi-modal learning interfaces make education more accessible: The future of AI in K12 could also see a shift toward multi-modal learning, where AI systems can process and interpret data from various inputs like voice, images, and gestures—ideal for younger children or students with disabilities who may not yet be proficient in typing or reading. For example, an AI could assist students by understanding spoken commands, analysing drawings, or recognising physical gestures like pointing at objects. This development opens up new ways for children to engage with educational technology, making learning more accessible and interactive. Startups focusing on these multi-modal capabilities will lead the charge in creating inclusive and intuitive learning environments.
As AI technology continues to evolve, its role in K12 education will become increasingly significant. While the path to widespread adoption is marked with challenges—budget constraints, infrastructure limitations, teacher hesitancy, large and powerful incumbents—AI’s potential to transform learning experiences is undeniable. From automating administrative tasks to personalising student learning paths and even supporting emotional well-being, the possibilities are vast and worth chasing after.
As ever, if you'd like to discuss this work or you are a founder working on a solution in this space, we would love to hear from you, so please do get in touch with rs@brighteyevc.com or via the deck submission form on our homepage.
We are grateful to Sabrina Bukenya, from Stanford's Graduate School of Business, for her support on this project.
NEXT UP: AI in HE
Sources:
OECD report 2023:
https://www.oecd.org/en/publications/oecd-digital-education-outlook-2023_c74f03de-en.html
Frontline Education Report: https://thejournal.com/Articles/2024/07/25/Frontline-Education-K12-AI-Adoption.aspx?admgarea=Features1
HolonIQ 2024 EDU Survey:
https://newsletters.holoniq.com/ai-in-k12-europe-edtech-200/
Magic School’s post regarding their raise:
https://www.magicschool.ai/blog-posts/announcing-magicschools-15m-series-a-raise
Teacher spending:
https://explore-education-statistics.service.gov.uk/find-statistics/la-and-school-expenditure
https://www.oecd.org/en/publications/the-funding-of-school-education_9789264276147-en.html
Tech applications:
Source showing AI curricular development: https://link.springer.com/article/10.1007/s40593-023-00358-x
France AI ambitions:
https://www.info.gouv.fr/upload/media/content/0001/09/02cbcb40c3541390be391feb3d963a4126b12598.pdf
Comments