Choice boards look great on a bulletin board. Nine squares, pick three, everybody's engaged. The problem shows up later, at grading time, when you realize the poster option and the essay option and the podcast option aren't actually measuring the same thing - and now you're comparing apples to dioramas.
A choice board only works as differentiation if every single option assesses the *same standard*, just through a different format. If it doesn't, you haven't differentiated instruction - you've just handed out nine different assignments and called it one.
Start from the standard, not the format
Before you touch a template, write down the one sentence that has to be true no matter which square a student picks. For a unit on proportional relationships, that sentence might be: "Student can identify and justify whether a relationship is proportional using at least two representations." Every option on the board has to let a student demonstrate exactly that - a poster, a video, a written explanation, a slide deck - the container changes, the proof doesn't.
This is the step most choice boards skip, and it's the one AI is genuinely good at helping with. Paste your standard and ask something like:
- "Here's my learning standard: [paste it]. Generate 6 different product formats a student could use to demonstrate mastery of this exact standard, keeping the required evidence identical across all 6."
You'll get back options that share a core - which is what makes the board gradable with one rubric instead of six.
One rubric, not six
This is the part that saves you actual time. If every option has to prove the same thing, you can grade with a single rubric that names the required evidence in format-neutral language: "Clearly identifies whether the relationship is proportional" and "Justifies the answer using at least two representations" work whether the student built a poster or recorded a screencast. Ask AI to draft this rubric directly from your standard sentence, then read it once yourself to make sure it doesn't accidentally reward the format ("uses color effectively") instead of the content.
A board that actually differentiates
Once the core evidence is locked, the *format range* is where the real differentiation happens - and it should span more than "poster vs. essay." A board that gives real choice usually includes:
1. A low-writing option (annotated diagram, recorded explanation) for students who know the content but struggle to produce text.
2. A high-structure option (fill-in template, guided worksheet) for students who need scaffolding to organize their thinking.
3. An open-ended option (student-designed product) for students ready to work without a frame.
4. A collaborative option, if your standard supports it, for students who think out loud better than they think on paper.
You don't need nine squares. Four solid, genuinely different-in-cognitive-demand options beat nine cosmetic variations on the same worksheet.
Watch for the option that's secretly easier
The one trap that undoes a choice board fast: one option that quietly requires less thinking than the others, and every kid who wants an easy grade picks it. Before you post the board, do a five-minute gut check - could a student complete each option without ever touching the actual standard? If yes for any square, either cut it or add a specific requirement (a labeled diagram, three examples, one written justification) that forces the thinking back in.
Where a differentiation plan pulls this together
If you're building choice boards regularly, it helps to have the underlying differentiation logic written down once rather than reinvented per unit - our [differentiation plan generator](/differentiation-plan) will draft the tiered options and shared success criteria from your standard, so you're editing a starting draft instead of building the whole board from a blank page every time.
Choice isn't the goal. Choice that all points at the same evidence is. Get that part right and the board does what it's supposed to - give kids real say in how they show you what they know, without giving you six different things to grade by six different standards.
