Nobody Ships the Hard Part of Curriculum

Every curriculum I've ever seen invests heavily in the wrong 90 percent.

The artwork is gorgeous. The teaching script is airtight. The take-home craft has a supply list down to the number of pipe cleaners. And then, buried at the bottom of page four, sometimes in a smaller font like an afterthought, you get the part that actually determines whether the lesson worked: the small group discussion questions.

Usually three of them. Usually terrible.

The most expensive part of any curriculum to write well is the cheapest part to phone in. So almost everyone phones it in.

I've spent a lot of the last few years looking at what churches actually buy versus what actually lands on a Sunday morning, and this gap keeps showing up. Publishers pour their production budget into things you can photograph for a catalog. Discussion prompts don't photograph well. You can't put "good question about the parable of the sower" on a box cover. So the questions get written last, fast, by whoever's left, and they come out sounding like a reading comprehension quiz.

"What did Jesus do?" "How does that make you feel?" "What can we learn from this?"

Dead on arrival. I've watched a room of eighth graders answer all three in under 40 seconds and then stare at the volunteer like, now what.

Here's what I think most curriculum buyers don't fully price in when they compare options. The teaching content is the part a decent teacher can improvise around. If the script is thin, an experienced person fills the gap. But the discussion is the part that breaks in the room, in real time, in front of people, and it breaks in a way you can't recover from on the fly unless you're unusually good. When a question lands flat, the silence isn't neutral. It costs you the room's trust. People decide, quietly, that this isn't going anywhere, and they check out. Getting them back is expensive.

A weak teaching point loses a minute. A weak question loses the whole group.

And yet the economics of curriculum production almost guarantee the questions stay weak, because the market rewards what's visible. Churches evaluating materials flip through, see polished pages, robust leader guides, nice design, and reasonably assume the whole thing is that quality. It usually isn't. The design quality and the question quality live in completely separate parts of the org chart, and one of them has a much bigger budget.

I'll be honest about the limits of my own take here. I don't think publishers are lazy or cynical. Writing genuinely good discussion prompts is hard, slow, unglamorous work. A single question that actually opens a group up, that has no obvious right answer, that a 45-year-old and a 19-year-old can both engage from where they are, might take longer to write than three pages of teaching content. It doesn't scale the way production art does. So I understand why it gets shortchanged. I just don't think buyers should keep paying premium prices while pretending the premium reaches the part that matters most.

This is the piece I'm genuinely curious about with AI in curriculum, and I want to be careful not to oversell it. The thing AI is actually decent at, when it's pointed at the right problem, is generating a lot of question variations fast and then letting a human throw out the 90 percent that are junk. Not writing the perfect question. Generating enough raw material that a teacher can find the two or three that fit their specific room. That's a different job than "write the lesson." I've seen it produce prompts that are worse than a good publisher's and I've seen it produce prompts that are better than a rushed one's. The deciding factor is almost always the person editing, not the tool. Anyone selling it as a replacement for teacher judgment is selling something that doesn't exist yet.

What I'd push ministry leaders to do is boring and it works: next time you're evaluating any curriculum, paid or free, skip the artwork entirely and go straight to the discussion section. Read the questions cold. Ask whether a real person in your actual context would say something interesting in response, or whether they'd give the answer everybody already knows. If it's the second one, you're paying for packaging.

The hard part is the part nobody ships. Buy accordingly.