SKILL 2 OF 9 · AI FEATURES
Feedback design
Deciding what someone reads after they act: who it is for, how long it is, and the one thing it asks them to do next.
WHERE IT SHOWS UP · PICK A PROJECT
PROJECT 04The Artificial Leaf
THREE READERS
One rule per reader
A LEARNER · PROJECT 02
A TEACHER · PROJECT 03
A YOUNG PERSON · PROJECT 01
WHERE
Mid-game, in Animal Planet
The teacher dashboard
The Planner Board
LENGTH
Ten words or fewer
—
A tag and one plain line
VOICE
First person, strict on the score
Decisions, in classroom terms
Their own plan, spoken back to them
ASKS FOR
One fix
Challenge or support, next lesson
One action
PROJECT 02 · LEARNING METERFROM P. 9–14
Ten words that tell a learner what to fix
Feedback is read mid-game, so it had to be short, quick and honest. Ten words or fewer, in the first person, naming one fix, and strict on the score. False praise was treated as a risk and measured.
OPEN PROJECT 02 ▸PROJECT 01 · BEELINEFROM P. 3–8
Nudges on a young person’s plan
The conversational layer reads the Planner Board and speaks about what is on it. Each nudge is a tag, one plain line and one action. Suggested cards are marked ⟡, so a young person can tell their own plan from a suggestion.
Why this matters…Missing step!More routes available!
OPEN THE CONVERSATIONAL LAYER ▸
PROJECT 03 · JUMP MATH LABFROM P. 15–18
A dashboard a teacher can act on
Three rounds of prototypes, each answering one question: can a teacher read this and know what to do next lesson?
ROUND 1CountsRaw counts of activity.
ROUND 2ProgressProgress through the work.
ROUND 3 ✓DecisionsDecisions a teacher can act on in class.
TOOLS
Azure OpenAIGPT-3.5 and GPT-4Figma
RELATED SKILLS
WORK WITH MEAI features, from design to evaluationWhere AI helps, rubrics and test sets, feedback people can act on
ASK ME FOR A DAY RATE ▸