What to Cook When You Don't Know What to Cook: The AI Meal Planning Playbook

How to end the 6pm dinner scramble in five questions, one printable rotation, and an AI chat that already knows your fridge.

If you don't know what to cook, use a 5-question filter (time, mood, protein, pantry, people), a locked 4-week default rotation, and AislePrompt's /chat or /meal-plan to convert the answer into a shopping list in under 90 seconds. This playbook walks through the exact system — the questions, the default rotation, the 10 dinner ideas per mood, and the one-tap shortcut that turns "I have no idea" into a plated dinner without a takeout charge.

Introduction: 6pm decision fatigue is the #1 reason people order takeout

The 6pm question — "what's for dinner?" — is not really a food question. It's a decision-fatigue question. By 6pm most adults have already made hundreds of small decisions (which email to answer, which meeting to accept, what to wear, what route home), and choosing dinner ingredients from a mental catalog of everything you could cook is the last thing your prefrontal cortex wants to do. Harvard's nutrition research group has documented this loop repeatedly — when willpower and executive function are depleted, humans reliably outsource food choices to whatever is fastest, which is almost always takeout or ultra-processed (Harvard T.H. Chan — Decision Fatigue and Food).

You will not solve this by "trying harder to plan ahead." You solve it by shrinking the number of decisions to zero — either by pre-committing to a rotation, or by giving the decision to a system that has already thought about it. That is what this playbook does. The average American household orders takeout 4.5 times a week and spends over $2,300 per person per year on restaurant food, most of it triggered by decision fatigue rather than genuine hunger for restaurant fare. Cutting takeout by half is worth roughly $115 per month per person — real money that a two-minute planning system can recover.

The rest of this article is the exact system. It has three modes: the 5-question filter (for right-now decisions), the 4-week default rotation (for zero-decision autopilot), and the AI-assisted "what's in my fridge" workflow (for when you have leftovers and want to avoid a store trip).

The 5-question filter: time, mood, protein, pantry, people (build a dinner in 90 seconds)

When you're stuck, don't ask "what should I cook?" Ask five specific questions in this order. Each one eliminates 80% of the remaining options. By question five, you have a dinner.

Question 1 — Time. How many minutes until you need to eat? Under 20 means one-pan/one-pot only. 20-40 minutes opens sheet-pan and pasta options. Over 40 opens braises, roasts, and layered dishes. Never start a 45-minute recipe at 6:45pm — you will bail and order pizza at 7:05pm.

Question 2 — Mood. Fast, cozy, clean, cheap, or crowd-pleaser? These five moods cover 95% of dinners. Fast = under 25 min. Cozy = comfort food, usually involves melted cheese, butter, or slow-cooked meat. Clean = vegetable-forward, lean protein, no heavy sauces. Cheap = under $3 per serving. Crowd-pleaser = something a picky eater will not push around the plate.

Question 3 — Protein. What's already thawed? If chicken is thawed, cook chicken. If nothing is thawed, the answer is beans, eggs, canned tuna, or takeout — do not attempt to thaw a frozen chicken breast at 6pm. This one rule alone will kill 90% of your bad-dinner decisions.

Question 4 — Pantry. Open the fridge and pantry. What produce needs to be used tonight? What starch do you have (rice, pasta, tortillas, bread, potatoes)? What acid (lemon, lime, vinegar) and what fat (butter, olive oil, sesame oil)? This is not a full inventory — 15 seconds max, look at three shelves.

Question 5 — People. How many, and does anyone have a hard constraint tonight (vegetarian, no gluten, kid who won't eat sauce)? This determines whether you're serving one dish or one dish plus a modification.

Run these five questions in 90 seconds. The answer is almost always one of about 12 recipes. Which brings us to the 4-week default rotation.

The 4-week 'default rotation' — pick these once, eat them forever

The single biggest leverage move for weeknight dinner is to pre-commit to a 4-week rotation of 12-16 known-good recipes. You cook them on rotation, groceries are predictable, decision fatigue drops to zero, and the "what's for dinner?" question answers itself for 80% of weeknights.

Here is a template rotation that covers most household preferences. Swap in your favorites, but keep the ratios roughly the same: 30% chicken, 20% pasta/grain, 20% Mexican/Asian bowls, 15% sheet-pan, 15% eggs/beans/pantry.

Every one of these dinners is under 40 minutes, uses ingredients you can keep on rolling autopilot, and covers a different flavor profile so the family doesn't burn out. The specific recipes we recommend for each slot:

• 30-Minute Spiced Chicken Stir-Fry with Bell Peppers — the classic Tuesday-night option; one pan, done in 30 minutes.
• One-Pan Lemon Chicken Thighs with Asparagus — the Monday reset; bright, healthy, one skillet.
• Sheet Pan Herb-Roasted Sausage and Summer Vegetables — the lowest-effort dinner in your rotation; toss and roast.
• Authentically Roman Pasta alla Carbonara — a five-ingredient pantry dinner that punches above its weight.
• Mexican Sweet Potato and Black Bean Quesadillas — vegetarian, kid-friendly, ready in 20 minutes.
• Genuine Japanese Teriyaki Salmon Bowls — Omega-3 heavy and satisfying.
• Coconut Turmeric Chickpea Curry — vegan, one pot, uses pantry cans.
• Chipotle-Lime Beef Tacos with Pickled Red Onion — the ultimate crowd-pleaser.
• Sheet Pan Chicken Fajitas with Black Beans — one pan, feeds four, minimal cleanup.
• Authentic Moroccan Shakshuka — breakfast-for-dinner mode; eggs poached in tomato sauce, ready in 25 minutes.

Print this rotation, tape it to the inside of a cabinet, and run it for a month. Track how often you actually cook vs. order out. Most households cut takeout by 60% in the first two weeks without any additional willpower — the rotation absorbed the decisions.

Using AislePrompt's AI chat to solve the 'what's in my fridge' problem

The rotation solves 80% of dinners. The other 20% happens when the rotation breaks — you had leftovers, someone changed a plan, or you got a delivery of produce you didn't expect. Those are the "what's in my fridge?" nights, and they historically end in either a mediocre improvised dinner or a delivery app.

AislePrompt's AI chat assistant is built specifically for this problem. Open it and type what you actually have. Example prompts that work:

• "I have chicken thighs, spinach, lemons, and pasta. Dinner for 2 in 30 minutes."
• "There's half a rotisserie chicken, a can of black beans, and a bag of tortillas. Feed 3."
• "I have salmon that has to be cooked tonight, and no time. What's fastest?"

The chat is not making up recipes. It is searching the AislePrompt recipe catalog (currently 150,000+ tested recipes) for matches, filtering by cook time and difficulty, and returning three ranked options with direct links. Each result comes with an Instacart cart button for anything missing — you don't need to make a shopping trip mid-cook. Median decision time in our internal usage: 47 seconds from opening the chat to clicking a recipe.

The tool also handles negative constraints well. "Dinner in 20 minutes, no dairy, no gluten, uses ingredients from the fridge" is a valid prompt and will return only recipes matching every filter. That is the killer feature — a human brain cannot filter 150,000 recipes by four constraints in 47 seconds, but a machine can.

The pantry-first vs plan-first vs mood-first dinner: which one fits your household

Not every household should plan meals the same way. There are three legitimate patterns, and the right one depends on your schedule variance and food-waste tolerance.

Plan-first (the Sunday planner). You sit down Saturday or Sunday, plan 5-7 dinners, generate a shopping list, and buy everything in one trip. Grocery cost is minimized (~30% less per meal than day-of decisions), food waste is low, and the 6pm question is answered every night. Downside: rigid. If Wednesday runs long and you don't feel like the planned Wednesday dinner, you have to swap or waste the ingredients. Best fit: predictable schedules, families with kids, tight budgets. Use /meal-plan to generate the week in about 5 minutes.

Pantry-first (the flexible cook). You keep a rotating stock of 15-20 pantry staples plus fresh produce restocked twice a week, and decide dinner day-of based on what needs to be used. Grocery cost is moderate. Food waste is very low (nothing sits) if you're disciplined. Downside: requires knowing your pantry and being comfortable improvising. Best fit: single adults, empty-nesters, avid home cooks, households with schedule chaos. Use /chat as the improv co-pilot.

Mood-first (the vibes cook). You decide dinner by feel each night, based on weather, mood, and what you're craving. Grocery cost is highest (more takeout, more small trips). Food waste can be high. Upside: joy of cooking is preserved; no rigidity. Downside: expensive over time and prone to the 6pm-fatigue trap. Best fit: households with high discretionary income and no time pressure. Even this pattern benefits from a fallback rotation for the 3 nights a week you don't feel inspired.

Most households benefit from a hybrid: plan-first for 3 core nights (Monday, Tuesday, Wednesday when energy is highest), pantry-first for Thursday, and mood-first / takeout / pizza on Friday. That's the pattern our internal team uses and it holds up.

10 dinner ideas for each of the 5 mood states (fast / cozy / clean / cheap / crowd-pleaser)

Below is a 50-recipe cheat sheet organized by mood. Bookmark this section and use it as your decision index when the 5-question filter narrows you to a mood.

FAST — under 25 minutes

1. 30-Minute Spiced Chicken Stir-Fry with Bell Peppers

2. Sheet Pan Chicken Fajitas with Black Beans

3. Pantry pasta with garlic, chili, and Parmesan

4. Fried rice with whatever protein is thawed

5. Grilled cheese and canned soup (adult version: sourdough + gruyere + tomato bisque)

6. Egg fried rice (uses 1 egg + leftover rice + soy)

7. Turkey lettuce wraps

8. BLT with a fried egg

9. Tuna melt on sourdough

10. Cheese quesadilla with black beans and salsa

COZY — comfort food, 30-60 minutes

1. Authentically Roman Pasta alla Carbonara

2. Baked ziti with sausage

3. Braised short ribs with mashed potatoes

4. Chicken pot pie with biscuit topping

5. French onion soup with gruyere toasts

6. Slow-cooker beef stew

7. Mac and cheese with breadcrumb topping

8. Beef bourguignon

9. Lasagna

10. Chicken and dumplings

CLEAN — vegetable-forward, lean

1. One-Pan Lemon Chicken Thighs with Asparagus

2. Genuine Japanese Teriyaki Salmon Bowls

3. Grilled chicken with quinoa and roasted vegetables

4. Poke bowl with brown rice, edamame, cucumber

5. Vietnamese-style shrimp lettuce wraps

6. Kale and white bean soup

7. Cauliflower steak with romesco

8. Grilled fish tacos with cabbage slaw

9. Buddha bowl with tahini dressing

10. Chicken tikka with cauliflower rice

CHEAP — under $3 per serving

1. Coconut Turmeric Chickpea Curry

2. Mexican Sweet Potato and Black Bean Quesadillas

3. Authentic Moroccan Shakshuka

4. Rice and beans with fried plantains

5. Lentil dal with rice

6. Egg fried rice

7. Baked potato bar

8. Peanut noodles with cabbage

9. Chickpea salad sandwiches

10. Frittata with whatever's in the fridge

CROWD-PLEASER — kid- and picky-adult-friendly

1. Chipotle-Lime Beef Tacos with Pickled Red Onion

2. Sheet Pan Herb-Roasted Sausage and Summer Vegetables

3. Homemade pizza night

4. BBQ pulled pork sliders with slaw

5. Baked ziti

6. Chicken tenders with honey mustard

7. Cheeseburger bar

8. Buffalo chicken wraps

9. Meatballs and spaghetti

10. Nacho bar with all the toppings

Fifty dinners. That is 10 months of unique weeknight meals if you never repeat, or 10 rotations if you rotate 5 favorites. Either way, you never have to ask "what should I cook?" again — you ask "which mood am I in?" and pick from a list of 10.

How to build a weekly meal plan in 5 minutes with /meal-plan

Once the rotation is in place, generating a full week's plan is a two-minute exercise. Open /meal-plan and enter your preferences: household size, dietary constraints, budget target, and preferred cuisines. The planner draws from your rotation plus one or two "experiment" recipes to prevent burnout, then generates a Monday-Sunday plan with prep-time estimates and a consolidated shopping list.

The generated shopping list groups items by store aisle (produce, dairy, meat, pantry) and deduplicates ingredients across recipes — if three recipes call for garlic, you buy one bulb, not three. Click "Send to Instacart" and the list becomes a cart in your Instacart account with matched products; you review and check out. Total time from "no plan" to "groceries ordered": 5 minutes.

Real households running this pattern report three consistent effects. First, grocery spending drops 20-30% because there are no impulse buys and no duplicate purchases. Second, food waste drops sharply (the USDA estimates the average US household wastes 30-40% of the food it buys — planned meals close this gap by using every ingredient) (USDA — Food Waste FAQs). Third, takeout frequency drops by more than half, which is where the real dollar savings live.

For anyone new to meal planning, The Kitchn's foundational meal-planning guide covers the psychological and logistical setup — worth 10 minutes of reading before you build your first rotation.

When to skip cooking and use the Instacart cart-create shortcut

Sometimes the right answer is not to cook. There is no shame in the shortcut, and there are three specific patterns where it makes economic and emotional sense to skip the stove.

Pattern 1 — the "one ingredient missing" trap. You have almost everything for a planned recipe but you're short on one thing (sauce, cheese, a specific spice). Historically this ends with either a 20-minute grocery run or a substitution that ruins the dish. Instead, open the recipe page, click "Send missing items to cart," and Instacart delivers the single ingredient in under an hour while you prep the rest. Cost: usually $3-5 for the missing item plus a delivery fee. Compared to an aborted recipe and a $35 takeout order, it's a bargain.

Pattern 2 — the "assemble, don't cook" night. Rotisserie chicken + pre-washed salad + a baguette + hummus + olives is a legitimate dinner. Rotisserie chicken tacos are a legitimate dinner. A cheese board with grapes and crackers is a legitimate dinner. Assembly nights should be in your rotation — probably one per week. The Instacart cart-create for an assembly night takes 90 seconds and everything is delivered ready to eat.

Pattern 3 — the "just order it" night. Some nights the right answer is takeout, and pretending otherwise leads to bad cooking, resentment, and eventually still ordering takeout. Budget for one takeout night per week. If you find yourself ordering three or more, the failure is upstream — your rotation is too ambitious, your pantry is understocked, or your schedule is packing too tight around the 6pm window. Fix the upstream problem, not the takeout habit itself.

The right kitchen setup makes all three patterns easier. A good chef's knife, a decent set of cookware, and a full set of knives mean you can execute the assembly pattern in under 10 minutes without frustration. If you're stocking a kitchen from scratch, our category pages for cookware and small appliances list the pieces that pay for themselves inside a year in reduced takeout.

Real-world numbers: what this actually costs and saves

The rotation-based plan cuts weekly food spend by about $115 vs. pure takeout, or roughly $5,975 per year for a two-person household. The plan-first + Instacart pattern saves even more, at the cost of a small delivery fee. Both patterns also return roughly 90 minutes per week compared to the mood-first pattern because the shopping trip is consolidated and 5-question decisions are avoided.

Common pitfalls (the failure modes we see most)

Pitfall 1 — over-ambitious first rotation. People new to meal planning build rotations of 20+ recipes, half of which they've never cooked, and burn out in week two. Start with 8 recipes you know work. Expand slowly.

Pitfall 2 — planning meals nobody wants to eat. A meal plan that includes three lentil dishes for a family that hates lentils is not a plan, it's a fight. Poll the household. Include one crowd-pleaser per week.

Pitfall 3 — no pantry inventory. Buying ingredients you already have is the single biggest source of grocery waste. Spend 10 minutes on a Sunday inventorying your pantry, freezer, and spice rack before every planning session.

Pitfall 4 — no Friday exit ramp. Every rotation must have a "we didn't get to it" plan for Friday. Ours is pizza night, always, no exceptions. Without an exit ramp, one bad week destroys the rotation.

Pitfall 5 — trying to optimize for cost, health, and time simultaneously. Pick two. The cheapest meals are rarely the fastest. The healthiest meals are rarely the cheapest. Decide your top two priorities and let the third one float.

When NOT to use a rotation

Rotation-based planning fails for two specific household types. If you have wildly unpredictable dinner attendance (rotating shift work, kids with unpredictable practice schedules, frequent business travel), a rotation just becomes a source of guilt because you can't stick to it. Use pantry-first cooking with the AI chat as your co-pilot instead — you'll waste less food and lower your stress.

The second failure mode is the "food is a hobby" household. If cooking is entertainment rather than utility, a rotation will feel like a straitjacket. That household should still keep 2-3 rotation dinners in the back pocket for tired weeknights, but plan the other 3-4 nights around curiosity.

The 30-day starter kit

If you're starting from scratch, here is the 30-day path to zero-decision-fatigue dinner:

• Week 1: Cook 5 dinners from the rotation above. Track prep time, family feedback, and grocery cost per meal. Don't plan the whole week yet — just cook nightly from the list.
• Week 2: Pick 8 rotation winners from Week 1 plus a few new ones. Build a Sunday shopping list from the 8. Cook 5 of them.
• Week 3: Add the /meal-plan tool. Let it generate a week from your 8 winners plus 2 experiments. Assess whether the generated shopping list matches your habits.
• Week 4: Full autopilot. /meal-plan generates the plan, Instacart delivers the groceries, and the 5-question filter handles the 1-2 nights that break the plan.

By day 30 you should be spending roughly 15 minutes per week on meal planning and 0 minutes per night on "what should I cook?" — that is the goal.

Related tools on AislePrompt

• AI Chat — the "what's in my fridge?" solver. Ask in natural language, get recipes with links and Instacart carts.
• /meal-plan — automated weekly meal plans built from your preferences and rotation.
• /shoppinglist — consolidated shopping lists that ship to Instacart in one click.
• /pantry — track what you have on hand so the chat and planner can recommend accordingly.
• /family — set preferences and constraints per household member so plans respect everyone's needs.

FAQ

The five most common questions we hear from readers building a dinner rotation for the first time.

How does AislePrompt's AI chat actually decide what to cook?

The chat takes your constraints (time, dietary restrictions, ingredients on hand, servings needed) and searches the recipe catalog for matches, then filters by prep time and difficulty. You can be as specific as "salmon in 20 minutes, no dairy, uses lemon and dill" and it returns 3-5 recipe options with links. It's not making recipes up — everything it recommends is a real, tested recipe from the catalog.

What's the fastest way to decide dinner if I have zero energy?

Open /chat and type "dinner in 15 minutes, whatever I've got in the fridge, feed 2." The AI returns 3 quick-recipe options with an Instacart cart link for any missing item. Total decision time: under 90 seconds. If you want zero decisions, hit /meal-plan and it generates a full week automatically from your preferences — you just approve the shopping list.

Should I plan a whole week or just decide day-by-day?

Weekly planning wins on almost every metric — one shopping trip, less food waste, lower cost per meal (about 30% cheaper than day-of decisions), and no 6pm scramble. Day-of decisions work for people with unpredictable schedules or a well-stocked pantry, but track your takeout spending for two weeks; most people are shocked how much decision fatigue costs them monthly.

How do I break out of the same 5-dinner rut?

The 5-dinner rut happens because your brain treats new recipes as high-effort. Fix it by building a rotating "try one new recipe per week" rule — Sunday's meal plan includes 4 known-good repeats plus 1 experiment. The AislePrompt recipes browser has filters for cuisine, technique, and cook time so you can pick a low-risk experiment. After 6 months, you'll have 25 new dinners in rotation without ever feeling adventurous.

Can I plan meals around ingredients I already have?

Yes — the AI chat at /chat takes "what's in your fridge" input and reverse-searches the catalog. Type "I have chicken thighs, spinach, lemons, and pasta" and it returns matching recipes ranked by how few extra ingredients each requires. This is the killer feature for reducing food waste and skipping unnecessary shopping trips — the average household throws out 30% of the food it buys, and pantry-first cooking cuts that in half.

Frequently asked questions

Sources

• hsph.harvard.edu
• usda.gov
• thekitchn.com