Almost all academy trusts are using AI in some way, with just 6 per cent reporting that they are not, according to the Confederation of School Trusts’ latest survey.
And a significant number of trusts are using AI to improve their special educational needs and disabilities support.
This is likely to become an area that more trusts start to look at amid reforms to the SEND system and as the government is urging schools to use AI to cut teacher workload, within certain guardrails.
Here are three projects that multi-academy trusts are developing to help their SEND teams and pupils with SEND:
1. Bourne Education Trust
Bourne Education Trust (BET), which runs 31 schools across Surrey, Hampshire and South London, is exploring AI technology to reduce pressure on staff and families.
It is building and piloting an AI assistant to support SEND teams with the administrative requirements of their roles, specifically to assist decision making.
Chris Goodall, data and digital innovation director at the trust, answered ձ’ questions about the project.
What problem is the project trying to solve?
“To cut the heavy administrative burden on SEND teams, meaning skilled staff spend more time with children rather than on paperwork.”
What does it involve?
“It has a very specific purpose, with functionality limited and clearly defined. It streamlines and organises a child’s SEND record, produces structured caseload summaries and helps to complete form fields from documents already held in school.
“The intention is that, if successful, the assistant will be available for use in other MATs.”
What were the main challenges and how were they overcome?
“The greatest challenge has been helping staff to understand that the tool is not a finished product, unlike most products purchased by schools. Staff are, in effect, beta-testing it alongside us, so their feedback is central to its development.
“The second challenge is ethical. The teacher must own the process from start to finish and take full responsibility for the outcome, with the AI acting as a support to professional judgement rather than a replacement for it.”
How much money and resource was spent on it?
“There are no real costs in development currently, as the work is being done as part of the collaborative approach across the Open Education AI community. Trusts have different options of how they could eventually build and access it, so it’s not easy to put a figure on it.”
How did you measure success and what was the outcome?
“At this early stage of development, accuracy is our primary measure. Alongside that, we are assessing usefulness: whether staff can see genuine value in the tool and how it supports their work in practice.”
What do staff think about it?
“Sendcos within the trust report that the assistant is releasing time, meaning they have more capacity to dedicate to provision for pupils.
“We have built a feedback loop into the project from the outset so that each round of use directly shapes the next iteration.”
2. Learning Academy Partnership
Learning Academy Partnership runs 22 academies in the South West.
One of the trust’s Sendcos has led the development of a bespoke AI agent to support target setting in individual support plans (ISPs), which have been proposed as part of the government’s SEND reforms.
Vlog talked to Emma Bone, deputy trust lead at Learning Academy Partnership, to find out more.
What problem is the project trying to solve?
“We recognised that teachers and Sendcos can sometimes find it challenging to create targets that are both specific and measurable, while also linking clearly to curriculum progression and SEND provision.
“One of the key challenges it addresses is that writing precise, measurable and achievable targets can be difficult and time-consuming, with approaches sometimes varying between teachers and schools.
” This tool aims to reduce paperwork, avoids repetition and ensures everyone has access to the same specialist thinking.”
What does it involve?
“Users receive well-structured draft targets that are measurable, teachable and consistent with our agreed approach.
“The agent acts like a knowledgeable colleague, sharing expertise, modelling best practice and freeing up time so teachers can focus on supporting pupils.”
What were the main challenges and how were they overcome?
“One of the main challenges is being really tight with the prompts for the agent. Another challenge we anticipate is making sure that all practitioners have necessary professional development to empower them.”
How much money and resource was spent on it?
“The trust uses a pay-as-you-go model for its AI agent development work, meaning costs are based on consumption, so it only pays for what is used.
“We currently have a number of agents running across our schools, at a combined cost of around £30 per month.”
How did you measure success and what was the outcome?
“Success has been assessed through trialling the tool in several schools, gathering feedback from teaching staff and refining it in response.
“Training has also been delivered to Sendcos across the trust to support its implementation and ensure that staff understand how to use it effectively.”
What do staff think about it?
“Teaching staff have responded positively to the potential for the tool to make writing ISP targets easier, particularly at the end of a busy term when teachers are balancing competing demands.”
3. Creative Education Trust
Creative Education Trust (CET) has 17 academies in the South East, Midlands and East of England.
The trust has developed its own AI agent, AdaptEd, to support adaptive teaching in the classroom and suggest practical strategies for teaching different learners.
The idea originated at The Hart School in Staffordshire, where vice principal Louisa Burnett wanted to find a new way to help teachers respond effectively, in real time, to the different needs of 30 young people in a classroom.
Ash Mudaliar, director of IT and edtech at CET, explained how it is working.
What problem is it trying to solve?
“AdaptEd was created to address what we describe as the ‘timing problem’ in adaptive teaching: teachers are often required to make rapid pedagogical decisions about different learners, needs and barriers in a matter of seconds, while managing the wider demands of the classroom.”
What does it involve?
“Teachers can use AdaptEd both when planning and in the moment during lessons. They give it some context about the learner, class or barrier they are trying to address, and it suggests practical adaptive teaching strategies they can consider.
“That might include different ways to explain or structure a task, support attention or regulation, scaffold learning, address misconceptions or help a pupil access an activity while keeping expectations ambitious.
“It is designed to support the teacher’s professional judgement, not replace it. The teacher decides which suggestions are appropriate for their learners and classroom context.”
What were the main challenges and how were they overcome?
“One of the main challenges was ensuring AdaptEd did not simply behave like a generic chatbot. We wanted its responses to reflect CET’s approaches to teaching, adaptive practice and inclusion, rather than generate completely open-ended suggestions.
“A second challenge was usability. We did not want teachers to have to become expert prompt writers before they could get value from the tool, so we created a simple front end that allows common information to be entered quickly.
“Teachers can still interact with it as a conventional AI assistant if they prefer.”
How much money and resource was spent on it?
“AdaptEd was developed in-house using Copilot Studio, with around 15 to 20 hours of development time, including engagement with quality of education and adaptive teaching/SEND specialists within the trust.
“It’s mostly about internal staff time, including training time for teachers. There is a Copilot licence cost, but it is relatively nominal, especially as it is also being used for other purposes as well.”
How did you measure success and what was the outcome?
“In the trial period, success was measured through teacher feedback, practical classroom examples, specialist review and quality assurance of the responses produced by the tool.
“The clearest outcome of the trial was that, following testing and refinement, AdaptEd was rolled out across all CET schools. We see this as an ongoing process, however, and will continue to use teacher feedback and classroom experience to improve the tool.”
What do staff think about it?
“We received formal feedback from 60 teachers, alongside a further 29 responses gathered during early testing at our inclusion conference.
“The most commonly reported impact was practical classroom support, with most respondents saying AdaptEd either gave them strategies they could use immediately or helped them adapt lessons more effectively. Teachers also reported increased confidence in supporting learners with SEND, and time savings.”
