Responsible AI in Voyage
AI that supports professional judgement.
Voyage helps staff find potentially relevant curriculum statements from a written observation. Staff use their knowledge of the child and the evidence in front of them to decide what to record.
- AI-assisted curriculum suggestions
- Staff make every decision
- Configured for in-region inference in AWS London
A useful suggestion, not an automated judgement
Find relevant assessment statements without losing the observation
Suggestions stay within the familiar assessment workflow, ready for staff to review alongside the observation.
-
Record the observation
A member of staff writes what the pupil did, said, made or showed in a learning record.
-
Surface possibilities
When staff request suggestions, Voyage ranks relevant framework sections and may highlight statements as they open those sections.
-
Make the decision
Staff review the candidates, decide what the evidence supports, choose the qualifier and save the record.
A clear route through the system
What the model request contains
Voyage sends curriculum-suggestion requests to Claude Sonnet 4.6 through Amazon Bedrock.
Sent for this task
- The written observation
- Relevant framework names and descriptions
- Relevant framework statements and examples
Not sent for this task
- Pupil name or identifier
- Photographs or video
- The wider learning record
- The pupil's plan
- Previous assessment history
Our approach
Designed around professional judgement
Purposeful assistance
AI is used to make it easier to find relevant parts of a curriculum framework from a written observation.
Professional judgement
Suggestions sit within the normal assessment workflow, where staff consider the evidence and make the judgement.
Clear data handling
We explain what is included in the model request, which service processes it and where that processing takes place.
Questions schools ask
How does AI help staff in Voyage?
Voyage can rank potentially relevant curriculum sections and statements from a written observation, helping staff move through a framework more efficiently.
Who reviews the suggestions?
The member of staff reviews the observation and any suggested sections or statements, chooses the appropriate qualifier and decides what to save.
What information is used to make suggestions?
The model request contains the written observation and relevant framework names, descriptions, statements and examples. The data route above lists what is and is not included.
Which service powers the feature?
Voyage uses Claude Sonnet 4.6 through Amazon Bedrock, configured for in-region inference in AWS London. AWS states that Bedrock inputs and outputs are not used to train base models or made available to model providers.
Can we discuss this as part of our review?
Yes. We can talk through the current workflow and data route with school leaders, SEND teams, data-protection leads and IT colleagues.
Bring us your questions
If you are reviewing Voyage for your school, we can talk through the workflow, data route and the information your procurement or data-protection review needs.