Adapting a text so a pupil can access it — without dumbing down the actual learning — is one of the most valuable and time-consuming things inclusion teams do. AI is genuinely good at this.
The distinction that matters most
There's a world of difference between lowering reading demand and lowering cognitive demand. A pupil with dyslexia may have excellent comprehension and reasoning but struggle to decode the text. Simplify the wrong thing and you remove the learning; simplify the right thing and you remove the barrier.
The instruction to include in every prompt: "reduce the reading demand without reducing the cognitive challenge or removing key content." Without it, AI produces a simpler and less demanding text — which fails the pupil. Naming the distinction explicitly is the single most useful technique here.
The core adaptation prompt
Prompt — reading age adaptation
Rewrite this text for a reading age of approximately 9, keeping all the key content and the same conceptual demand. Use shorter sentences, more common vocabulary and clearer structure, but do not remove information or simplify the ideas themselves. Keep any subject-specific terms and briefly explain each: [paste text].
The instruction to keep and explain subject vocabulary is important — pupils with SEND still need the technical terms; they just need support accessing them.
Producing multiple versions at once
Prompt — tiered versions
Produce three versions of this text at reading ages of approximately 7, 9 and 11. All three must contain the same key information and support the same learning objective: [paste text].
Other adaptations worth knowing
Prompt — dyslexia-friendly structure
Reformat this text to be more accessible for a pupil with dyslexia: shorter paragraphs, clear subheadings, key information in bullet points where appropriate, and no long unbroken blocks. Keep all content: [paste].
Prompt — vocabulary pre-teaching
From this text, identify the 8 words most likely to be a barrier for a pupil with a reading age of 8, and write a child-friendly definition and example sentence for each: [paste].
Prompt — comprehension scaffolding
Write comprehension questions on this text at three levels: retrieval, inference and evaluation. Keep the language of the questions simple even where the thinking required is demanding: [paste].
Tools that help
Any general AI assistant handles these prompts well — see our comparison of AI tools for teachers. Some school platforms have built-in reading-level adaptation, which is faster for routine use but less flexible than a well-written prompt. Start with prompts; add tools if volume justifies it.
Always check the output
AI occasionally changes meaning while simplifying, drops a crucial qualifier, or introduces Americanised vocabulary. Read the adapted version against the original before it reaches a pupil — particularly where accuracy matters, as in science or history.
And keep the pupil's dignity in mind
Adapted materials should look like the same task, not visibly like an easier one. Ask for versions with the same title, layout and structure — a child who receives an obviously different sheet knows exactly what it means. See personalising learning with AI.
Want your inclusion team adapting resources in minutes?
We train SEND and inclusion staff on the prompts that work. AskColin gives schools low-cost monthly support with one-to-one help whenever you need it — so when you're stuck on a real job at 8am, there's someone to ask. Practical, jargon-free, built around your team.
Get one-to-one AI support for your schoolFrequently asked questions
How can AI adapt texts for SEND pupils?
Ask it to rewrite for a specific reading age while keeping all key content and the same conceptual demand — shorter sentences, commoner vocabulary and clearer structure, without simplifying the ideas. Always instruct it to keep and briefly explain subject-specific terms.
How do I lower reading demand without lowering challenge?
Include the explicit instruction: 'reduce the reading demand without reducing the cognitive challenge or removing key content.' Without it, AI produces a text that is both simpler and less demanding, which removes the learning rather than the barrier.
What other SEND adaptations can AI do?
Reformatting for dyslexia with shorter paragraphs and clear structure, identifying vocabulary likely to be a barrier and writing child-friendly definitions, and producing tiered versions of the same text at several reading ages that all support the same objective.
What should be checked in AI-adapted texts?
AI sometimes changes meaning while simplifying, drops crucial qualifiers, or introduces Americanised vocabulary. Read the adapted version against the original before it reaches a pupil, especially where accuracy matters. Also keep the layout similar so adapted work doesn't look visibly easier.