By EdTech, we mean pedagogical programmes and approaches that use technology to support educators and learners with teaching and learning. This may include computer adaptive learning and Generative Artificial Intelligence (GenAI) such as large language models (LLMs).
GenAI has become widely used by the general public in recent years: for example, large language models (LLMs) such as ChatGPT, released in November 2022. Consequently, GenAI is being increasingly used in teaching and learning: in a 2025 National Literacy Trust report, 2 in 3 (66.5%) of 13 – 18 year olds reported using AI to support literacy and learning
EdTech can be used to support different approaches to teaching and learning. Some common uses within the learning process might be: summarising information; elaboration or explanation, guided reasoning or modelling solutions, providing formative feedback, metacognitive support to guide students through planning, monitoring and evaluating their learning and providing and supporting personalisation and differentiation. Approaches could be used for the whole class or targeted for specific pupil groups. The EEF are further developing the evidence base to better understand which learner cognitive processes are worth protecting from AI-enabled offloading and which could be supported by AI, to help schools and educators in making better-informed pedagogical choices.
We have several completed and live projects which use technology to personalise learning: for example, Maths Whizz and Lexia Reading Core5®, which offer adaptive learning approaches based on the level of content mastery of individual pupils. Our ChatGPT in lesson preparation Teacher Choices trial supports the generation of robust evidence on the impact of Generative AI on teacher workload.
Launched in May 2026, a new DfE-EEF EdTech Testbed aims to continue to build a stronger evidence pipeline to help schools and colleges make more informed decisions about which EdTech approaches and products to buy and adopt, including products that use GenAI. The Testbed programme includes two stages:
- Stage 1: Up to 100 rapid evaluations in schools and colleges. The DfE will commission a research and delivery partner to lead this stage, with the EEF acting as an‘evidence partner’ by providing methodological advice and ensuring alignment with high-evaluation standards.
- Stage 2: Large-scale robust evaluations of programmes with the highest promise. The EEF will deliver up to 15 randomised controlled trials, or other robust evaluations (QEDs), of potentially high-impact tools identified through Stage 1.
The programme will focus on generating evidence about EdTech approaches that support key priorities across the sector, including:
- reducing workload
- improving learner outcomes
- increasing inclusivity for all learners, including those with SEND.
Why it matters to our mission
EdTech, particularly products using GenAI, is an area where more research is needed to understand the impact on learning, including for economically disadvantaged learners. As AI use is increasingly prevalent in schools and settings, it is important that we continue to build the evidence base. AI products and programmes are often highly scalable, so if found to have evidence of positive impact, there is the possibility for a programme to be scaled to support large numbers of learners.
EEF evidence looking at which elements of EdTech programmes and approaches may lead to impact, and how these findings apply to socio-economically disadvantaged children and young people, suggests that the effectiveness of EdTech interventions is best characterised as mixed. The impact of EdTech on disadvantaged pupils is even smaller and less certain, meaning that without careful implementation it risks widening rather than narrowing the disadvantage gap.
Previous EEF-commissioned evidence and practice reviews have found that there is limited robust evidence about what works in EdTech. Teachers are increasingly expected to integrate technology into their classroom practices, often linked to improving pupil attainment, yet the evidence base presents mixed messages about when and how to use technology in the classroom. This results in a lack of guidance on what practices actually work when using EdTech to enhance pupil learning outcomes.
The EEF is working to build the evidence base, with the recent publication of the meta-analysis for the systematic review of EdTech interventions for disadvantaged pupils. This comprehensive review identified the following mechanisms of EdTech interventions that are associated with improved pupil attainment outcomes:
- Grounding technology in pedagogical processes;
- Grounding technology in inclusion and accessibility;
- Applying diverse media for learning;
- Creating dynamic feedback mechanisms;
- Leveraging technology for adaptive teaching;
- Planning for implementation;
- Guiding the development of reflective and personalised environments;
- Managing game-based learning and gamification;
- Scaffolding collaborative learning.
We are undertaking further work as part of a project funded by Google.org to support educators to access the evidence on EdTech, including an open-source tool for teachers with guidance on the most effective ways to use technology and AI, based on evidence from global trials and evaluations.
Please see also the EEF’s initial evidence review of Digital Technology (2012) and the EEF’s guidance report on Using Digital Technology to Improve Learning (2021), which is designed to support senior leaders and teachers to make better informed decisions based on the best available evidence we currently have.
| EEF Toolkit | The Teaching and Learning Toolkit incorporates approaches involving EdTech within the 30 Toolkit strands. As and where there are relevant findings from digital technology approaches, these are noted in the ‘Key findings’ and ‘Behind the average’ summaries. |
| Evaluation | We have several completed and live projects that use technology to personalise learning: Lexia Reading Core5® has a promising result on pupil attainment from a research trial. Our ChatGPT in lesson preparation Teacher Choices trial supports the generation of robust evidence for the use of Generative AI in teachers’ activity and lesson preparation, looking at the impact on teacher workload. The trial generated promising findings with teachers using ChatGPT for lesson preparation identifying a 31% time saving per week compared to those not using GenAI. A recent trial of Maths Whizz, an intelligent tutoring programme, showed evidence of positive impact for Key Stage 2 learners (age 6 to 10 years old), and greater potential impact for economically disadvantaged leaners. |
| Sector Engagement | Stakeholder engagement helps to shape contextual understanding and we welcome further opportunities for collaboration. EdTech is an area of governmental interest as an opportunity for innovation in education. Policy initiatives stem from the increased use of EdTech in schools and settings and continued expansion of the EdTech market meaning that support is needed for educators to be able to make pedagogical choices about EdTech that is both safe and impactful. |
In 2023 we included EdTech as a specific funding theme, focusing on established education programmes already able to deliver to approximately 50+ settings, delivered to pupils aged 5 – 16 (Key Stage 1 – 4) in England. The scope of this funding round was identified via stakeholder consultation, including a large-scale practitioner survey, and reviewing existing research evidence. The EEF commissioned the following programme evaluations: Maths-Whizz, Reading Plus, FFT Tutoring with the Lightning Squad and the Teacher Choices Chat GPT trial.These programmes use EdTech in some or all of the following ways:
- formative assessment;
- to support the development of specific skills (for example, English, maths);
- computer adaptive learning or AI technologies for teaching and learning.
EEF subsequently funded a second rapid Teacher Choices trial, to assess Oak National Academy’s AI-powered teaching support product, Aila, across time saving for teacher lesson planning. Results are due in September 2026.
We plan to commission further evaluations looking at the use of GenAI by learners age 16 – 19 years. Recent funding rounds have included a focus on how 16 – 19 learners and educators are using AI-based tools to support providing feedback, reducing workload and professional development for practitioners and supporting socio-economically disadvantaged learners particularly.
In 2026, we launched a new invitation to tender for up to four innovative research projects to look at the impact of GenAI on learner cognition on learners age 13 – 15 years. The EEF is especially interested in how GenAI may lead learners to “offload” thinking tasks like recall, planning, reasoning, or drafting onto GenAI tools. We expect to finalise commissioning from Autumn 2026.
EdTech trials and pilot studies in progress
| Priority area | Highest-level EEF project | Project | Delivery team | Year groups | Level of intervention | Description | Headline results |
|---|---|---|---|---|---|---|---|
| EdTech | Effectiveness | Lexia Reading Core5® | Queen’s University Belfast and Lexia | Year 2 | Targeted, whole class | A trial to test the impact of Lexia Reading Core5® (Lexia), a computer-based integrated learning system that aims to improve reading skills by providing children with individualised reading instruction and practice in six areas: phonological awareness, phonics, structural awareness, automaticity, fluency, vocabulary, comprehension. | The evaluation report will be published in Summer 2026. Efficacy trial: +1 month 4 padlocks |
| EdTech (16−19) | Pilot | STEER Tracking to support 16+ Transition, Attendance and Retention | STEER Education | Year 11 and 12 | Whole class | A pilot trial of STEER Tracking: a whole-class, socioemotional programme that aims to reduce the risks of disengagement and absenteeism by giving school-based sixth forms and colleges early-warning data on learner wellbeing. | The evaluation report will be publishing in Spring 2028. |
| EdTech | Efficacy | FFT Tutoring with the Lightning Squad | FFT Education | Year 3 and Year 4 | Targeted, whole class | FFT Tutoring with the Lightning Squad is a targeted online reading catch-up programme that aims to improve reading skills for pupils aged 7 – 9 in Years 3 – 4 who are below age-expected reading levels. The programme is delivered by FFT Education, an established provider of educational programmes and services to schools. | The evaluation report will be published in Spring 2027. |
| EdTech | Efficacy trial | Reading Plus (2024/25 trial) | Reading Solutions UK | Year 5 | Targeted | Trial to test the impact of Reading Plus, a whole class online adaptive EdTech reading programme, on reading fluency and reading comprehension. | The evaluation report will be published in September 2026. |
EdTech Choices evaluations in progress
Priority area | Highest-level EEF project | Project | Delivery team | Year groups | Level of intervention | Description | Headline results |
|---|---|---|---|---|---|---|---|
| EdTech | Teacher Choices | Lesson planning using AI lesson assistant, Aila | NFER, Oak National Academy | Year 3 – 6 | Key Stage 2 teachers (generalist) | Teacher Choices trial testing the impact of using Aila on lesson preparation time and resource quality for Key Stage 2 class instruction. | The evaluation report will be published in Autumn 2026. |
Completed trials in both EdTech and associated priority areas
| Priority area | Highest-level EEF project | Project | Delivery team | Age (yrs) | Level of intervention | Description | Headline results |
|---|---|---|---|---|---|---|---|
| EdTech | Efficacy | Maths-Whizz Intelligent Tutoring | Whizz Education | Year 2 – 5 | Targeted, whole class | Trial to test the impact of Maths-Whizz, a whole class online intelligent tutoring programme and implementation planning programme, that aims to improve maths outcomes. | +1 months’ progress 4 padlocks |
| Teacher Choices, EdTech | Teacher Choices | ChatGPT in lesson preparation | Bain & Company, Hg Foundation, NFER | Year 7 – 9 | Whole class, targeted | Teacher Choices trial exploring different approaches to lesson and resource preparation in Key Stage 3 science. Specifically, this trial tests the impact of ChatGPT-assisted lesson and resource preparation on teacher time, with implementation support via a guide for teachers, against approaches unassisted by generative AI technology. | Delivery completed, 30% reduction in teacher planning time. |
| EdTech | Effectiveness | Stop and Think: Learning Counterintuitive Concepts | Behavioural Insights Team | Year 3 and Year 5 | Targeted, whole class | A project to test a software programme that uses quizzes and games to help pupils learn counterintuitive concepts in science and maths. | Delivery completed, discussing next steps and considered a Promising Programme due to its positive findings in science attainment: +2 months 3 padlocks Efficacy trial: +2 months 4 padlocks |
| Reading | Effectiveness | Accelerated Reader | NFER and Renaissance Learning | Year 7 | Targeted, whole class | Testing the impact of a web-based programme that encourages reading for pleasure and helps track progress. | 0 months 5 padlocks Efficacy trial: +3 months 3 padlocks |
| Reading | Effectiveness | Abracadabra (ABRA) | Coventry University | Key Stage 1 | Targeted, whole class | Testing an online and a paper-based version of a balanced approach to reading. | Effectiveness: +2 months 4 padlocks Efficacy trial: +3 months 4 padlocks |
| Tutoring | Effectiveness | Affordable Maths Tuition | NESTA | Key Stage 2 | Targeted, whole class | Testing the impact of online maths tutoring. | 0 months 3 padlocks |
| Literacy | Effectiveness | Units of Sound | Dyslexia Action | Year 7 | Targeted, whole class | A computer-based phonics programme for Year 7 pupils who are struggling with reading. | −1 months 1 padlock |
| Personalised learning | Efficacy | Flipped Learning | Shirelands Collegiate Academy | Key Stage 2 | Targeted, whole class | Pupils undertake some learning prior to lessons, freeing up class time for feedback and personalised support. | +1 month 3 padlocks |
| Literacy | Efficacy | GraphoGame Rime | University of Cambridge | Year 1 and Year 2 | Targeted, whole class | Testing a computer programme designed to improve pupils’ literacy through teaching phonics via “rhyme analogy”. | −1 month 5 padlocks |
| Literacy | Efficacy | Headsprout Early Reading in Special Schools | Bangor University | Key Stage1 and 2 | Whole school | A targeted literacy programme aiming to build fluency in essential early reading skills for children with SEND in special schools. | 0 months 1 padlock |
| Subject skills | Efficacy | onebillion | University of Nottingham | Key Stage 1 | Targeted, whole class | Maths learning app monitored by teaching assistants. | +3 months 5 padlocks |
| Literacy | Efficacy | Tutoring with Alphie | Success for All | Key Stage 2 (Year 6) | Targeted | Tutoring with Alphie is a computer-assisted programme that aims to improve the literacy skills of struggling readers. | +2 months 0 padlocks |