BETTER TEACHER FEEDBACK WITH AI
Designing an AI-assisted feedback experience to support teacher development and improve the quality of classroom feedback.
Quick Facts
1
Background
A school-led organisation dedicated to improving teacher training and professional development across England wanted to provide an alternative feedback method for Early Career Teachers (ECTs) carrying out lessons in classrooms. A secure lesson recording tool integrated into the tools used by ECTs was built, enabling offline observations and mentor feedback with AI.
2
Skill Areas
-
User Research
-
User Journey Mapping
-
Usability Testing
-
Design Strategy
-
Ideation facilitation
-
Feature Prioritisation
-
Product Design - Prototyping
-
Cross functional collaboration with UCD and technical team members
-
Agile
3
My Deliverables
-
Research synthesis and playback
-
Design principles
-
Figma designs for two user journeys ready for engineering handover
-
Feature prioritisation collaborated with cross-functional team
-
Lesson recording tool 'How to' guides
-
Tool training and onboarding
The Details
THE CHALLENGE
Early Career Teachers (ECTs) rely on high-quality feedback to reflect on their teaching and develop their practice. But for mentors (who are also teachers) observing and supporting them, capturing meaningful feedback from a lesson can be time-consuming and difficult to structure consistently.
The challenge was to explore how AI could help teachers capture their lessons for observation and turn these recordings into useful, actionable feedback - reducing the administrative burden of observations while keeping mentors firmly in control of the professional judgement and development process.
CLIENT
School led organisation
YEAR
2026
INDUSTRY
Public Sector
KEY GOAL
Design an AI-assisted experience that makes it easier for teachers to capture, review and provide meaningful feedback on lessons, helping ECTs get more consistent support to improve their practice.
MY ROLE
I was the delivery team Researcher and Designer in charge of designing two journeys - for ECTs and mentors, testing the interaction of designs and working closely with tech colleagues to build out the tool. Following build I also drove the tool onboarding and upskilling activities with schools so the team could get further insights on tool improvements.
Some of the key deliverables within my role included:
-
Research synthesis and playback
-
Design principles
-
Figma designs for two user journeys ready for engineering handover
-
Feature prioritisation collaborated with cross-functional team
-
Lesson recording tool 'How to' guides
-
Tool training and onboarding
UNDERSTANDING THE USER GROUP
Although both ECTs and mentors are teachers, they interact with the feedback process from very different perspectives and have different needs from the tool for their feedback workflow.
Early Career Teachers (ECTs) are developing their teaching practice and rely on feedback to understand what is working well, identify areas for improvement and build confidence in the classroom. For them, feedback needs to be clear, constructive and actionable - helping them translate observations into meaningful changes in their practice. With strict timings on observations (due to mentor capacity), ECTs are unable to get the breadth of feedback that they'd ideally like.
Mentors, on the other hand are experienced teachers responsible for observing, supporting and guiding ECTs. Alongside their own teaching responsibilities, they need to identify meaningful moments from lessons, reflect on what they have observed and provide feedback that is specific enough to support an ECTs development. This makes the feedback process an additional demand on their time and cognitive capacity.
This created two distinct needs:
1. ECTs needed feedback that was useful for their development
2. Mentors needed a more efficient way to product high-quality feedback without compromising their professional judgement.
Understanding this distinction became central to exploring where AI could support the process - not by replacing the mentor's expertise, but by reducing the effort involved in turning lesson observations into meaningful feedback.
BREAKING DOWN THE PROCESS
A test and learn approach with user-centred design was used to deliver screen designs quickly so feedback could be captured before the tool went live. We aligned with client stakeholders on the first iteration of the tool based on the most value for ECTs, mentors and tech feasibility.
-
Ran user research to uncover pain points and prioritised user requirements for a lesson recording tool. User journeys for our target groups - ECTs and mentors were mapped with opportunities prioritised for the first iteration of the tool.
-
Co-designed the features that could be available within the lesson recording tool through a facilitated cross-functional team.
-
Prototyped wireframes for research and testing with ECTs and mentors.
-
Revised wireframes informed by research, stakeholder requirements and feasibility.
-
Design documentation for engineering handover.
-
Collaboration with engineering colleagues to break up features into build tickets.
-
Onboarding and upskilling of a small group of initial schools to use and test built tool.
TITLE OF THE CALLOUT BLOCK
THE IMPACT
-
The recording tool and AI labelling platform for feedback has been rolled out to 30 mentors with over 1000 users due to be onboarded to the tool in September.
-
The tool is continuing to develop with features being prioritised as learnings are gathered in roll out.
-
ECTs are already seeing that this tool is helping with feedback from mentors noting that it is 'eliminating vague feedback'.