If you have not touched accessibility settings since September, you are not alone. The biggest shift is not one headline feature, but the steady improvement of built-in tools becoming “good enough” for everyday classroom use. Live captions are more reliable across devices, dictation has improved in accuracy and punctuation, and reading supports are less clunky to turn on quickly. That matters for SEND because the barrier is often not whether a feature exists, but whether it is predictable, fast and easy to normalise.
The second change is the way AI is appearing inside familiar platforms. Instead of sending pupils to a separate tool, more AI support now lives where they write, read, search and organise. That can be helpful, but it also raises the stakes for boundaries: when AI sits next to the work, it is easier for it to become the work. The goal this term is to treat AI as a scaffold that fades, not a crutch that replaces thinking.
A “minimum viable inclusion stack” is the smallest set of tools and routines you can rely on in any room, on any day, with minimal fuss. It is not a full SEND toolkit, and it is not a replacement for specialist provision. It is a baseline that reduces variability, so pupils are not constantly re-learning how to access learning.
In practice, your inclusion stack has three layers. First, universal access tools that everyone can use without stigma: captions, read-aloud, zoom, high contrast and focus modes. Second, personalised supports that can be switched on quickly for individuals: dictation, simplified reading views, word prediction and guided access. Third, classroom routines that make the tools usable: a consistent “access check” at the start of lessons, agreed file formats and a plan for when tech fails.
If you want AI in the stack, keep it narrow and purposeful. Choose a small number of teacher-controlled uses that reduce workload and improve clarity, and a small number of pupil-facing routines that build independence. For wider multimodal classroom moves, you may find it helpful to pair this with the four-channel classroom playbook so that text, audio, visuals and interaction are planned deliberately.
In Google Workspace classrooms, the fastest wins are usually in Chrome and Docs. Turn on Chrome’s built-in accessibility features where possible, and explicitly teach pupils how to use reading mode, zoom and high-contrast options. In Docs, voice typing can be a strong scaffold for pupils who can explain ideas orally but struggle with transcription. The key is to pair it with a short editing routine so spoken language is shaped into written language, rather than submitted raw.
For video and whole-class instruction, reliable captions reduce the cognitive load for pupils with auditory processing needs, attention differences, or EAL needs alongside SEND. If you are supporting multilingual learners too, it is worth reading AI for EAL beyond translation and applying the same principle: comprehension scaffolds first, language shortcuts last.
Microsoft’s strength remains integrated reading and writing supports. Immersive Reader can reduce visual clutter and help pupils control spacing, font and read-aloud pace. Dictation in Word and OneNote can support idea generation, while built-in Editor tools can be framed as “proofreading prompts” rather than “automatic correction”. In Teams, consider how captions and transcripts are stored and who can access them, especially where pupils discuss sensitive personal information.
A practical classroom move is to provide “two versions” of the same resource: the standard handout and an accessible version designed for Immersive Reader or a clean digital layout. This is not about lowering expectations; it is about removing avoidable barriers so the learning goal is reachable.
On iPads and Macs, the most term-ready approach is often to standardise a small set of settings: text size, display adjustments and spoken content. Guided Access can be valuable for pupils who need reduced distraction or a controlled app environment. For pupils who benefit from AAC, Apple’s ecosystem can be strong, but only if you plan how AAC use fits into classroom talk, turn-taking and assessment evidence.
Where teachers sometimes get stuck is consistency. If one room expects AirDrop, another expects email, and a third expects a learning platform upload, pupils who already struggle with executive function lose learning time. Your inclusion stack should include one agreed “handover method” for work and evidence.
Chromebooks remain popular because they are simple to manage and quick to start, which is a genuine accessibility advantage. The quick wins are in ensuring pupils know how to access built-in dictation, select-to-speak and display adjustments without needing an adult each time. If your school uses shared devices, consider how accessibility settings persist (or do not) between users, and whether pupils need a quick “access reset” routine at the start of each session.
Chromebooks are also a reminder that inclusion is not only about features; it is about friction. A tool that takes six clicks to enable is not a tool in a busy classroom. Wherever possible, reduce steps and make access part of the normal lesson start.
These micro-routines are designed to be brief, repeatable and clearly bounded. They work best when pupils know the rule: AI helps you start, check, or practise, but it does not replace your thinking.
First, use AI to generate three tiered explanations of the same concept (simple, standard, stretch) and then choose the best for a specific pupil profile. Second, ask AI to rewrite your instructions into a “Do/Check/Submit” sequence so pupils with executive function needs can follow steps. Third, create a vocabulary pre-teach list with child-friendly definitions and one example sentence, then teach it explicitly before the task.
Fourth, build a “worked example plus fading” sequence: ask AI for a model answer, then for a partially completed version with prompts, then for a blank template. Fifth, use AI to produce three hinge questions that test the same misconception in different ways, then use them in live checking. Sixth, create a short retrieval quiz with spaced repetition prompts, keeping questions tightly aligned to what was taught.
Seventh, during independent practice, let pupils use AI for “hint, not answer” prompts. A simple script is: “Ask for one hint, then show your next step.” Eighth, use AI to turn a dense text into a structured outline, but require pupils to add evidence quotes themselves, so comprehension remains active. Ninth, use AI to generate sentence starters matched to the genre, then remove them gradually over weeks as pupils internalise structure.
Tenth, for feedback, ask AI to convert your success criteria into a pupil-friendly checklist, then have pupils self-assess before you mark. This keeps the human judgement where it belongs, while AI reduces the time spent rewriting criteria.
If you want a planning-friendly way to embed these routines across subjects, adapt the lesson moves in AI across the curriculum and keep the SEND lens explicit: which barrier are we removing, and what independence are we building?
Ready to Revolutionise Your Teaching Experience?
Discover the power of Automated Education by joining out community of educators who are reclaiming their time whilst enriching their classrooms. With our intuitive platform, you can automate administrative tasks, personalise student learning, and engage with your class like never before.
Don’t let administrative tasks overshadow your passion for teaching. Sign up today and transform your educational environment with Automated Education.
The most common problem is not that tools do not work, but that they compete. Captions can clash with screen readers if both speak at once. Dictation can be undermined by aggressive autocorrect. Overlays and reading rulers can be lost when a web page reflows. AAC users can be excluded if AI-powered “chat” becomes the dominant participation route.
Start by deciding which tool is primary for each pupil in each context. If a pupil uses dictation, agree when it is used (idea capture, first draft) and when it is not (spelling practice, handwriting goals). If a pupil relies on text-to-speech, ensure AI tools do not replace reading practice with summaries too early; summaries can be a scaffold after first exposure, not instead of it.
Exam access arrangements need particular care. If a pupil is entitled to a reader, scribe, or word processor, your everyday tools should mirror that support without drifting into unauthorised assistance. A safe rule is to treat generative AI as “help with process, not product”: planning prompts, clarification and checking are safer than generating content that is assessed.
When you are evaluating an AI tool for SEND use, evidence and compatibility matter more than marketing. Ask what the tool demonstrably improves, for which learners, and under what conditions. If a vendor cannot explain limitations clearly, that is a risk signal.
A simple procurement checklist should cover data protection, bias and human oversight. Where is data stored, and is pupil data used to train models? Can you turn off retention? Does the tool allow pseudonymous use? For bias, ask how the system performs with non-standard grammar, dialect, AAC output, or speech differences, and what happens when it is wrong. For transparency, can staff see what prompts were used and what outputs were generated, so decisions are auditable?
Compatibility is often overlooked. Check whether the tool works with screen readers, keyboard navigation, captions and dictation. Confirm it functions on the devices you actually have, including older hardware, and that it does not require constant account switching. Finally, insist on a human-in-the-loop design: staff must be able to review, edit and override outputs easily.
If you are unsure how to evaluate vendor claims, use a structured approach like the classroom evaluation protocol in this article and apply it to SEND scenarios, not just general attainment.
Some “helpful” uses of AI are not appropriate for SEND contexts. Do not encourage pupils to disclose personal diagnoses, home situations, or mental health details to a chatbot, even if the tool claims to be safe. Do not use AI as a behaviour mediator (“Tell the bot why you were angry”), because it shifts pastoral responsibility away from trained adults and may store sensitive data. Do not rely on AI to determine need, label pupils, or suggest provision; that is a professional judgement requiring evidence and context.
Safer alternatives keep adults in the loop. Use AI to simplify the language of a school policy, not to interpret a pupil’s personal circumstances. Use AI to draft a neutral email template, then personalise it carefully. Use AI to create practice questions, not to generate assessed answers. When pupils use AI, set a clear boundary: if the output influences their final work, they must be able to explain what they accepted and why.
A two-week pilot is long enough to learn, but short enough to stop if it creates problems. In week one, choose one year group or teaching team and agree the inclusion stack settings you will standardise. Run a short “access check” routine at the start of lessons: captions on/off, text size, read-aloud availability, and how pupils will request help discreetly. Introduce only two AI micro-routines, such as instruction reformatting and hint-not-answer prompts, and track whether they reduce repeated questions and increase task completion.
In week two, add two more micro-routines and test assistive-tech compatibility in real lessons. Collect pupil voice in a structured way: what helped, what got in the way, and what felt unfair or confusing. Keep staff notes focused on observable impact, not novelty. At the end, decide what becomes standard, what remains optional, and what needs further safeguarding review.
For staff briefing points, keep it simple: inclusion stack first, AI second; scaffolds must fade; adults remain accountable; and no tool is introduced without a plan for data, access and failure modes. For a one-page parent/carer note, explain the purpose in plain language: built-in accessibility helps pupils access learning, and AI will be used in limited, teacher-guided ways to support planning, clarity and practice. State what data is (and is not) shared, how pupils are supervised, and how families can ask questions or opt out where appropriate.
May your inclusion routines become second nature, and your scaffolds quietly raise every pupil’s independence.
The Automated Education Team
Classroom displays can be powerful teaching tools, but only when they are …
This ready-to-run INSET Day workshop helps staff move from general AI …
Timetabling and daily cover are high-stakes operational puzzles, but AI is …
An AI Acceptable Use Policy written once and filed away won’t keep pace …
2025–26 brings sharper, more practical expectations for how schools manage …
A back-to-school AI plan succeeds when it is small, predictable and safe …
Clearing moves fast, but students deserve decisions that are calm, …
Results season can feel like a rush of numbers, narratives and urgent …
Summer reading can slip quietly, especially for pupils who need extra …
See õppeperioodiks valmis uuendus koondab selle, mis on päriselt uus (ja kasulik) sisseehitatud …
Den här terminsklara uppdateringen samlar det som faktiskt är nytt (och användbart) i inbyggd …
Tämä lukukauteen valmis päivitys kokoaa yhteen sen, mikä on aidosti uutta (ja hyödyllistä) …