How we meet the DfE's AI product safety standards

Last updated: October 2026

The Department for Education publishes "Generative AI: product safety standards" for schools and suppliers in England. It was last updated in January 2026.

There is no DfE approval or certificate for these standards. This page sets out how Assistant SENDCo meets them, so you can check for yourself.

What the product is for

Assistant SENDCo helps school staff write, review and share SEN Support Plans under the SEND Code of Practice. It is used by class teachers, SENDCos and school leaders, in schools from nursery to Year 11.

It is a staff-only tool. Pupils do not have accounts, do not log in and do not use the AI.

Our AI features fall under four of the DfE's use cases:

  • Administrative and management: suggested wording for targets and support in a plan.
  • Personalised learning and accessibility: provision ideas for a pupil's area of need, chosen by staff.
  • Content creation and delivery: ideas for adapting a lesson, and printable sheets a teacher can check and print.
  • Digital assistant: an in-app helper that answers questions about using the app.

Standards for products pupils use directly

Five standards cover products that pupils use themselves: filtering, monitoring, cognitive development, emotional and social development, and mental health. They do not apply, because pupils never use our AI.

Where AI helps make a printable sheet for pupils, a teacher reads it before it is printed, and the AI is told not to include any names.

Security

  • Every account uses two-step sign-in.
  • Staff accounts have set roles. Only the SENDCo can switch AI off for the whole school.
  • The AI is told to answer in fixed formats, and every answer is checked before it is shown.
  • It is instructed to treat what staff type as information, never as instructions.
  • We update the software regularly and run automated safety tests every day.
  • Progress 55 is certified to Cyber Essentials.

Privacy and data protection

  • AI is off until each member of staff turns it on, and a school can turn it off for everyone.
  • No pupil name and nothing typed into a plan is sent to the AI. It receives the statements and choices a teacher picks on screen.
  • In two places a teacher types to the AI: describing a lesson, and asking the helper a question. The app keeps pupils' names out on the teacher's device before anything is sent.
  • Our AI provider does not use this data to train its models.
  • Your school is the data controller; we are the data processor, with an independent Data Protection Officer. A data protection impact assessment is available to schools.
  • Our privacy notice and "Use of AI" page explain what is sent, where it goes and why.

Intellectual property

Everything in your plans belongs to your school. We claim no rights over it, and it is never used to train AI.

Design and testing

  • Every AI suggestion is a starting point. Nothing is added to a plan until a teacher chooses to add it.
  • The AI is told never to name or suggest a medical condition for a child. We check every answer for this and block any we find.
  • Our AI features are covered by automated tests that check no pupil details are sent.
  • Before we change the AI model behind a feature, we run safety tests on the helper, lesson help and provision ideas, and keep the results.

Governance and complaints

  • You can report an AI suggestion from inside the app, including each provision idea.
  • A new report alerts us straight away, and a person reads every report.
  • Our "Report a concern" page has a route for inappropriate AI content and for safeguarding, data protection and security concerns.

No manipulation

  • We run no adverts and no tracking tools.
  • We do not count or compare how much individual teachers use the app.
  • There are no streaks, rewards or pressure to keep using it.
  • The AI is told not to invent problems with work that is already good.
  • The AI is told never to claim or imply that it is a person or a member of staff.

Read more