How I Stopped Dreading the
5pm "What's for Dinner?" Question

It's 5:12pm. You've just walked in the door. The kids are already in the kitchen asking what's for dinner. You've got nothing defrosted, half a bag of pasta, some sad-looking zucchini, and a brain that's been making decisions since 7am and has absolutely nothing left.

You know this feeling. It's not hunger — it's dread. The specific low-grade panic of a question you have to answer right now, when you have the least capacity to answer it well.

That was my life for years. And I want to tell you how I fixed it — not with a 3-hour Sunday prep session, not with a $30/month meal kit subscription, and not with a fancy app. With a 10-minute conversation with an AI, once a week.

The 5pm feeling you know too well

There's a specific flavour of exhaustion that hits around dinner time when you haven't planned. It's not just tiredness. It's the weight of having to make one more decision — a big one — when your decision-making tank is completely empty.

For me, it was Tuesday evenings that broke me. Monday I could usually manage. By Tuesday the adrenaline of a new week was gone, there was nothing in the fridge that went together, and the options were: spend 20 minutes figuring out what to cook from a random assortment of ingredients, order takeaway we couldn't really afford, or make spaghetti bolognese for the fourth time that fortnight.

The thing that made it worse was that I knew, logically, that meal planning would fix it. I just couldn't make meal planning stick.

What made it worse (the guilt, the rotation)

We had a rotation. Seven or eight dinners that the kids would reliably eat, that I could cook without thinking too hard. It worked in the sense that we ate dinner every night. But my husband started calling it "the wheel" — as in, "what's on the wheel tonight?" And the kids knew the wheel. They'd call it before I said anything.

"Is it taco night or pasta night?"

Funny the first time. Less funny by the hundredth.

The guilt about it was real too. I genuinely wanted to feed my family real, varied, nourishing food. I had the knowledge. I just didn't have the bandwidth to translate that knowledge into dinner at 5pm when I was already running on empty.

What didn't help: apps, boards, and Sunday prep

I tried things. Of course I tried things.

I downloaded meal planning apps. The good ones required me to manually build a library of recipes first, which was its own overwhelming task. The ones with pre-loaded recipes were full of meals that weren't quite right — too many packaged sauces, too fussy, not what my kids would touch.

I had a Pinterest board with hundreds of saved recipes. It became a place where good intentions went to die. Scrolling through it at 5pm to find something I had ingredients for was worse than starting from scratch.

I tried Sunday meal prep — the proper kind, where you roast a tray of veg and cook a big pot of grains and portion everything into containers. I did it three Sundays in a row and felt amazing about myself. By week four, Sunday had other things in it, and the whole system collapsed.

The problem with all of these approaches was the same: they required a significant investment of time and mental effort upfront, and they didn't adapt to real life — different weeks, different budgets, different energy levels.

The first time I used AI for meal planning

I was sceptical. I'd used ChatGPT for other things and found it generic. The kind of output that looks helpful but isn't actually useful for your specific situation.

But I was also desperate. It was a Sunday evening, I had nothing planned for the week, and I thought — fine, let's see what it does.

I typed something like: "Plan 5 dinners for a family of 4 this week. My kids are 7 and 9. We want real food, no packet sauces. Budget around $120 for the week. My 7-year-old won't eat mushrooms or anything too spicy."

It gave me five dinners. Real ones — honey garlic chicken thighs, a lentil soup with crusty bread, a simple fish with roasted veg, homemade pasta with a basic tomato sauce, and a frittata. All meals I'd actually make. All meals my kids would eat. Prep times were accurate. It even flagged which meals I could partially prep ahead.

I was surprised. Not blown away — it was still just a list of dinners — but genuinely surprised that it had nailed what I asked for on the first try.

Then I asked it to generate the grocery list. It came back organised by section — produce, protein, pantry, dairy — with quantities. I cross-checked against my fridge and removed what I already had. The whole thing took about 12 minutes.

The moment it clicked

The magic didn't happen that Sunday. It happened on Tuesday evening.

It was 5:15pm. Kids were in the kitchen. And instead of that familiar drop in my stomach — the "oh god, what are we having" — I just... knew. Honey garlic chicken thighs. I'd bought everything on Sunday. It was already marinating in the fridge because I'd had 3 minutes on Monday morning and chucked it in a bag.

There was no decision to make. Tuesday evening required zero mental effort. I just cooked.

That felt significant. Not just convenient — actually significant. Because I realised the dread wasn't about the cooking. I don't mind cooking. The dread was about the deciding, every single evening, from scratch, when I had nothing left. Remove the decision, and the cooking is fine. Actually, the cooking is kind of enjoyable again.

The 2 prompts that changed everything

Here are the exact two prompts I use every Sunday. They're simple. They work.

The weekly dinner planner

This is the main one. I run this first, review the output, swap anything that doesn't feel right, and I'm done with dinner decisions for the week.

Plan 5 family dinners for a family of [size] for the week ahead.

Rules:
- Use whole, minimally processed ingredients — no packet sauces or artificial additives
- Each meal must be achievable in under 45 minutes on a weeknight
- Budget: under $[amount] total for the week
- [Child's name] won't eat: [list their specific exclusions]
- Include at least one easy meal for Wednesday or Thursday (midweek slump)

Output: meal name, main ingredients, prep time, any component I could do ahead.

The grocery list follow-through

Once I've confirmed the weekly plan, I paste it into this prompt and get a shopping list I can take straight to the supermarket — or paste into an online order.

Based on these 5 dinners: [paste your meal plan]

Generate a complete grocery list:
- Grouped by supermarket section (produce, protein, dairy, pantry)
- With quantities (e.g. "500g chicken thighs", not just "chicken")
- Flag anything I likely already have in a standard pantry
- Estimated cost per category if possible

That's it. Two prompts. Ten minutes. Week sorted.

For more detail on the full prompt system — including prompts for budget planning, batch cooking, and handling picky eaters — see our complete guide to AI meal planning. Or grab the free prompt pack with all of them in one place.

What Sunday looks like now

Before: Sunday had this low-level anxiety attached to it. I knew I should plan the week's dinners. I rarely did. And by Tuesday I was paying for it.

Now: Sunday takes about 10 minutes. I run the prompt while the kettle boils. I review the five meals, swap one if something doesn't feel right for the week ahead. I run the grocery list prompt. I check the pantry. Done.

I don't prep food on Sunday. I don't chop anything or cook ahead. Just the planning. The physical cooking still happens on weeknights — but it's execution, not problem-solving. I already know what I'm making. The ingredients are in the fridge. The decision is made.

The Sunday 10 minutes feels nothing like the Sunday "I should really plan dinners" anxiety of before. It's a small, satisfying task that closes a loop. And the rest of the week runs so much more smoothly because of it.

It's not about the food

I want to be clear about what actually changed here — because it's not really about having better dinners, though the dinners are better. It's about where the mental load sits.

Before, the mental load of dinner was distributed across the whole week. Every evening had a small but real cognitive cost: what are we having, do I have the ingredients, is there time, will the kids eat it. Multiplied across five evenings, that's a significant weight.

Now, that weight sits entirely in Sunday morning's 10 minutes. The weekday evenings are free of it. That's not a small thing. Decision fatigue is real. Reducing the daily decision-making burden has a genuine effect on how the rest of the evening feels — how patient I am, how present I am, how much energy I have left for the kids after dinner.

AI didn't fix my cooking. It fixed the mental load around my cooking. And that turned out to be what I actually needed.

If you want to go further — if you want a permanent household system where your budget, your family's preferences, and your whole food rules are locked in so you never have to re-paste them — Meal Planning OS is built exactly for that. But honestly? Start with the two prompts above. That alone will change your Tuesday evenings.

Frequently asked questions