AI Fluency Academy
DESIGN BRIEF R2 · DELIVERABLE 02
THE LESSON SYSTEM, IN DEPTH

The full anatomy of a lesson, where reading becomes capability.

Round 1 showed one lesson. Round 2 designs every state it moves through, opening, reading, example, practice, reflection, module check, and the resume state a returning learner meets. The practice task is the signature surface and is treated with the care Round 1 gave the homepage hero.

STATES  7
PRIORITY  Practice task
FIDELITY  Production · interactive
COMMITMENT  Practice is the lesson
ANATOMY OF ONE LESSON
A · Opening → B · Reading body → C · Inline example → D · Practice task ★ → E · Reflection → F · Module check ↺ G · Resume state
lesson / opening
Module 2 · Constrain, Lesson 3 of 6
Tell the model what good looks like
Vague requests get vague results. The fix is not a trick, it is a habit.
A

Lesson opening

How a lesson begins: a hook and a premise stated plainly. No "learning objectives" bullets, those belong to slide decks, not essays. The title sets the move, an italic standfirst makes a promise, and the reading starts immediately. The teacher's voice is present from the first line.

Commitment in play: "The teacher is present." We never write "in this section we will learn."
SCROLLS INTO
lesson / reading
When you ask an AI to "summarize this," you are leaving every decision to the model: how long, what to keep, what voice.
"You are not writing code. You are giving direction."
The fluent move is to state the shape of the answer before you ask for it.
AsideA note in the margin, where it earns its place.
B

Reading body

Dominant interaction

Long-form prose with editorial discipline. This is the dominant interaction, not the alternative to a video. Measure (~66 characters), rhythm, pull quotes, and asides in the margin where they earn their place. Typography decisions made here propagate to every lesson, so this screen sets the type system for the whole curriculum.

TYPE SPEC
Newsreader ~19px body, 1.7 line-height, 62–68 char measure, claret pull quotes.
DEVICES
Pull quote, margin aside, sparing bold. No callout-box clutter.
ANCHORED BY
lesson / example
See the difference one sentence makes:
VAGUE
"Summarize these customer interviews."
CLEAR
"Give me five recurring themes, each with a one-line quote. No preamble."
C

Inline example

The heart of teaching

The Vague / Clear comparison from Round 1, extended and made a recurring rhetorical pattern. Examples are not optional illustrations, they are the heart of the teaching. Every concept earns its place with at least one concrete before-and-after. The pattern propagates across every lesson so learners come to expect it.

Commitment in play: "Examples carry the work." Red/green is used only for vague/clear, never as decoration.
D

The practice task

★ HIGHEST PRIORITY · PRODUCTION FIDELITY

The single most important screen in the Academy, where reading turns into capability, and the screen that most shapes the credential's downstream pass rate. The learner tries the move themselves: an empty input, the task in plain language, a hint mechanism for when they get stuck, and an example answer they can reveal afterward, never before they've attempted it. This mockup is live, type, ask for a nudge, reveal the example.

aifluency.academy / foundations / 2-3 · practice
PRACTICE · CONSTRAIN

Your turn. Make a vague request clear.

Here is a request a colleague might fire off without thinking. Rewrite it so it names a format, a limit, and one exclusion, the three things you just read about.

THE VAGUE REQUEST
"Write up the notes from our team offsite."
T
Don't aim for perfect. Aim for clearer than the line above. That's the whole skill.
YOUR CLEARER VERSION{{ charCount }} characters
{{ draftStatus }}
NUDGE 1 · FORMAT
What shape should the write-up take? A summary paragraph, then bullets? A table of decisions and owners? Say so.
NUDGE 2 · LIMIT & EXCLUSION
Add a length ("under 200 words") and one thing to leave out ("nothing that wasn't actually decided").
ONE STRONG VERSION
"Turn the offsite notes into a one-page recap: a three-sentence summary, then decisions as bullets, then action items as a table with owner and due date. Under 250 words. Leave out anything we didn't actually decide."
Format, names the exact shape: summary, bullets, table.
Limit, "under 250 words" bounds the output.
Exclusion, "leave out anything we didn't decide" prevents invention.
EMPTY INPUT
No pre-filled text. The blank page is the point; the learner must produce the move.
HINTS, NOT ANSWERS
Two graduated nudges. They point at the structure without handing over the wording.
REVEAL AFTER
The model answer unlocks only after an attempt, and it is annotated, the teaching continues.
NO PASS / FAIL
Practice is never scored. "Clearer than the line above" is the only bar.
SETTLES INTO
lesson / reflection
Before you go, a question for you.
When would you reach for Constrain? When would you not?
Saved for you. We'll show it again when you return.
E

Reflection prompt

After the practice task, a short writing prompt the learner answers for themselves, "When would you use this move? When would you not?" Their answer is saved and surfaced again on the next visit, which is what makes the resume state feel personal rather than generic. It is never graded or shared.

PERSISTS
Stored per learner, re-shown in the resume card (state G) and the dashboard.
TONE
A teacher's question, optional to answer, never a required field.
AT MODULE END
lesson / module-check
Module 2 · Constrain, integration
Three short scenarios to knit it together
SCENARIO 1 · PICK THE MOVE
A teammate's output keeps drifting off-format. Which move fixes it, and how?
SCENARIO 2 · WRITE THE CONSTRAINT
F

Module check

Reinforcement, not gating

End-of-module integration: three or four short scenarios that knit the module's lessons together, pick the move, justify the choice, write the constraint. Not multiple choice. Not a quiz. A reflective check where the platform then shows the learner what a fluent answer looks like. A pass mark exists and is generous; the goal is reinforcement, not gating.

Same shape as the Capstone, smaller. The module check rehearses the comprehension pattern the Foundations Capstone (D3) completes.
↺ LEARNER LEAVES, THEN RETURNS
lesson / resume
Where you left off, six days ago.
LAST LESSON, IN ONE LINE
State the format before you ask, and the model stops guessing.
YOUR REFLECTION
"I'd use Constrain whenever the format matters more than the words…"
Continue, Lesson 4
G

Resume state

↺ Ties back to D1 state 05

What a learner sees when they return after a gap, the per-lesson companion to the journey-level resume in D1. The previous lesson's essence in one line, their own reflection from last time, and a generous "continue" affordance. It re-grounds without re-teaching, and never guilts the gap.

PULLS FROM
The saved reflection (state E) and last lesson's one-line essence.
DETAIL → D4
The dashboard-level lived-in view across day 1 / 30 / 90 is Deliverable 04.
DECISION EDoes the module check have a real pass gate?

The brief says a pass mark "exists and is generous; the goal is reinforcement, not gating." That leaves one genuine choice about what happens if a learner is below it.

E1
Never blocks, always advances
Below the mark just suggests a revisit. The learner can always continue. Purely reflective.
E2✓ CHOSEN
Soft gate, easy override
Nudges a retry but offers "continue anyway." A gentle speed bump, not a wall.
E3
Must pass to proceed
Hard gate. Contradicts "reinforcement, not gating", listed only for completeness.
AI FLUENCY ACADEMY · DESIGN BRIEF R2 · DELIVERABLE 02 OF 06 NEXT ↗ D3 · THE FOUNDATIONS CAPSTONE