Teachers make everyday choices about whether and how to use AI to support lesson planning. This project explored whether using Aila, Oak National Academy’s AI-powered lesson planning and resource creation tool, reduced teachers’ lesson preparation time while maintaining resource quality.
Teachers in Years 3 to 6 were randomly allocated to use or avoid Aila for lesson planning across all subjects for ten weeks during autumn term 2025. Both groups could use other generative AI tools but were asked not to find new resources on the Oak National Academy website. Researchers compared lesson preparation time and resource quality using weekly teacher diaries, surveys and independent teacher assessments.
Teachers increasingly need robust, accessible evidence on whether generative AI tools can reduce workload without compromising the quality of teaching or pupil outcomes. An earlier EEF trial found that Key Stage 3 (KS3) science teachers using ChatGPT saved 31% of their lesson preparation time. However, the trial tested a general-purpose AI tool rather than one specifically designed to support lesson planning and resource creation in line with the National Curriculum.
This project built on that evidence by testing Aila, a purpose-built AI tool designed to support lesson planning and resource creation, and exploring whether it could help teachers save time while maintaining the quality of their resources.
- Teachers using Aila spent 76% of the time on lesson and resource preparation compared with teachers not using Aila – around 2.5 hours per week, against 3.3 hours per week, a saving of roughly 49 minutes per week on average. This evaluation has a high security rating, reflecting a rigorous design, strong delivery, and confidence in this finding.
- This time saving did not come at the expense of quality. Analysis of the lesson resources produced showed that resources from the Aila group were no worse than those from the non-Aila group.
- Greater use was associated with greater time savings. Teachers who used Aila more often saved more time, and this pattern held up even when the first week and half-term break were excluded.
- Overall use of Aila was relatively low: 81% of teachers in the Aila group used it at least once, but it was used for only around 7% of lessons in an average week, and usage declined over the ten weeks. This may be related to a disconnect between Aila’s intended purpose of supporting ‘planning from scratch’ and the infrequency that teachers needed to do this. Teachers did not always need to plan whole lessons from scratch, which is Aila’s main function, and their existing planning habits and need to adapt the resources they downloaded from Aila meant they did not use it consistently. Even with this limited use, it was linked to a meaningful reduction in workload.
- Benefits were greatest for teachers newer to the profession or with lower confidence in their subject or pedagogical knowledge, or those with higher confidence in technology.
- Teachers had mixed views on Aila. Some found it accessible and aligned with the curriculum, and valued being able to use it as much or as little as they wanted. Teachers also found it particularly useful when planning lessons on unfamiliar topics. However, many preferred other AI tools, and fewer than one in five felt that using Aila had meaningfully changed their teaching or planning. In practice, teachers tended to use Aila alongside their existing planning processes, rather than as a replacement for them. In accordance with the intended use of Aila, many teachers adapted Aila’s outputs before they were ready to use in the classroom.
Almost all teachers in this trial reported lesson planning as a source of stress and perceived they spend too much time on it. This evaluation adds to encouraging evidence that generative AI tools can reduce the time teachers spend on lesson preparation.
Schools could consider generative AI tools as one way to reduce workload, particularly for routine tasks such as adapting materials, generating resources or planning unfamiliar content. However, time saved does not necessarily mean better planning or teaching: teachers still need to check and adapt AI-generated materials to ensure they are accurate, well sequenced and appropriate for their pupils and the wider curriculum. There should also be consideration about the extent to which particular AI tools encourage a complete cognitive offload of the lesson planning task, and the potential impact of this on professional expertise.
The findings suggest AI planning tools are most useful when teachers can opt to use them flexibly, according to their specific needs. Benefits were greatest for teachers newer to the profession or had lower confidence in their subject knowledge or pedagogical knowledge, suggesting these tools may be particularly valuable for teachers with less experience or confidence in planning.
The trial also suggests that providing an AI tool does not automatically lead to sustained use. Even when a tool is free and curriculum-aligned, teachers may use it only occasionally if they already have access to preferred tools and resources.
Schools and developers should view AI planning tools as a potential source of targeted support, rather than a replacement for teachers’ professional judgement or existing planning processes. Continued testing, refinement and adaptation are likely to be needed for these tools to reliably meet teachers’ needs.
Further research would help clarify what specific tasks teachers use Aila for.