Aila helped primary teachers cut lesson preparation time in an English trial

Teachers offered Aila reported about 49 fewer minutes of lesson and resource preparation per week in a ten-week English primary-school trial. The evaluation, released in October 2026, compared access to Aila with a group that could still use other AI tools. The saving applied to preparation for their class, rather than the teachers’ whole working week.[1]

The interviews explain why a measurable time saving did not translate into universal enthusiasm. Some teachers valued help with unfamiliar topics; others found the materials inflexible or needed substantial adaptation. Aila use declined during the trial. Its whole-lesson planning process fitted teachers starting from scratch more readily than those adapting an established bank of resources.[1]

A lesson starts with what pupils should learn

Aila’s design begins with a learning outcome and breaks the lesson into manageable learning cycles. Oak describes a sequence that connects what pupils already need to know with explanation, questions and practice. That structure gives the teacher several opportunities to inspect the educational substance before turning it into classroom materials.[2]

The curriculum shapes the generation

Oak’s technical explanation describes retrieval-augmented generation: the system brings relevant curriculum material into the context used to produce an answer. It can anchor a new lesson to an existing Oak lesson, drawing on a sequenced collection developed by teachers and subject specialists. That gives the generation a particular educational reference point.[3]

The prompts also contain guidance about the components of a lesson, with examples of suitable and unsuitable outputs. Teachers can request adaptations to reading level or local context. These are distinct design choices: retrieval supplies relevant material, while prompting directs how the system uses it. Neither removes the need to decide whether the finished explanation and tasks suit the pupils who will receive them.[3]

A practical place to begin

Oak’s current guidance describes Aila as a free beta tool with fair-use limits. A teacher needs an account and should use a computer or laptop; mobile use is not supported. It can produce editable lesson plans, slide decks and starter or exit quizzes. The option to revise and download those materials matters when a teacher wants help with one part of a lesson while retaining the rest of an established sequence.[4]

That selective approach also appeared in an earlier NFER experiment. In its 2024 trial with 259 Year 7 and Year 8 science teachers, the ChatGPT group commonly used AI for individual activities, such as questions and quizzes. Teachers reported using saved time for other planning and teaching work, or to reduce workload. The comparator in that trial was a group asked to avoid generative AI; its results therefore answer a different question from a trial in which other AI tools remain available.[5]

A sensible first use is a lesson component: a starter that checks prerequisite knowledge, a practice activity or a quiz that exposes a misconception. The teacher can examine whether the output serves that purpose, adapt it and keep the useful parts. Aila’s staged design makes those decisions visible throughout the planning process.[2]

Report an error or suggest a correction