AI dinner suggestions from your fridge

The problem families face

It's 4:30pm on a Wednesday. There's cooked rice from Monday, half a tin of coconut milk, a courgette that needs eating today, some frozen prawns, and a block of cheddar that's about to go dry at the edges. Every item is edible. None of it is obviously a dinner. Figuring out what combination makes a meal the family will actually eat — and getting it on the table by 6pm — requires a kind of low-grade creativity that most adults simply don't have left after a full day of work, commuting, and school pickups. So the courgette gets ignored, the prawns stay frozen, and someone orders Domino's.

The food waste problem and the 'what's for dinner?' problem are the same problem in different clothes. Households don't waste food because they don't care — they waste it because items purchased for a specific dinner that didn't happen fall out of the planning picture immediately. The courgette bought for Tuesday's pasta bake sat out because Tuesday was exhausting and you had cereal instead. By Thursday it's forgotten. By Saturday it's gone soft. The planning system never knew it existed.

• Converting whatever's in the fridge into a dinner the family will eat requires creative energy that's rarely available at 5pm
• Ingredients bought for a specific meal that got skipped fall out of the planning picture and get wasted
• Fridge contents are only known to whoever last did the shopping — no shared visibility across the household

Common ways families try to solve this today

The standard advice is to plan every dinner before you shop — a complete weekly menu with a precise ingredient list. This works well for the meals that are planned. It does nothing for the half-used tins, the produce that ripened faster than expected, or the leftovers from a kit service portion that was slightly too large. A complete weekly meal plan minimises food waste from planned purchases but doesn't touch waste from the gap between what was bought and what was eaten.

Searching 'dinner with courgette and prawns' on Google or a recipe site is the usual real-time fix. It finds results, but those results are optimised for the recipe writer's pantry rather than yours — they typically require three or four additional ingredients you don't have and won't use the rest of the tin or packet of something you buy specifically for one recipe. You end up at Tesco for one item while the original leftover still expires. The search engine doesn't know what you're trying to avoid buying.

• Weekly meal planning: reduces planned food waste but doesn't address mid-week surplus or unexpected items
• Recipe search engines: find relevant results but optimise for full recipes, not 'use what you have' constraints
• Meal kit services: pre-portioned so waste is minimal on kit nights, but don't cover the other half of the week

A better system for family planning

The operative principle is making the fridge contents a live input to the planning process, not a separate thing you check at the door. Food waste doesn't happen because people forget to open the fridge — it happens because what's in the fridge isn't visible at the moment the planning decision gets made. A system that knows what you have and actively surfaces meal suggestions based on what needs using first changes the dynamic: instead of discovering the courgette when it's already soft, you see it as a dinner suggestion on Monday when it's still perfect.

The second component is personalisation. Generic recipe search returns results for the ingredients in isolation. A system that knows your family — what the children will actually eat, what intolerances apply, what was popular last month — filters those results before presenting them. Three to five realistic suggestions for your specific household and your specific fridge contents is more useful than a hundred results ranked by SEO rather than by relevance to your evening.

• Make fridge contents a live input to dinner planning, visible at planning time not just at the fridge door
• Suggestions prioritised by what needs using first reduce waste without requiring manual tracking
• Personalisation filters output to meals the family will actually eat, not generic recipe matches

Example of a weekly system

After dinner or after the main shop: a quick confirmation of what was used and what remains. One to two minutes. The next afternoon, when someone opens the app and asks 'what should we have tonight?', the AI suggests three dinners using what's in the fridge, prioritised by shelf life. You pick one, and the shopping list generates for any missing ingredients — just the gaps, not the whole dinner. Sunday planning: the AI includes remaining fridge and freezer items in its weekly menu suggestions so the week starts with leftover stock already accounted for.

When the fridge is nearly empty: the AI sees the low stock and can generate a full week's meal plan with a corresponding shopping list — suggestions calibrated to your household's preferences, current season, and typical budget range. The same tool that suggests tonight's dinner from leftovers also plans next week's meals from scratch. It doesn't require switching modes — you ask the same question and the system responds to what's actually there.

• After dinner: confirm what was used to keep fridge inventory accurate for next day's suggestions
• Next afternoon: ask for dinner suggestions based on what needs using soonest
• Sunday planning: AI incorporates remaining fridge and freezer stock into the week's menu suggestions
• Near-empty fridge: AI generates a full meal plan and shopping list for the coming week

How Zenframe helps

Zenframe Assistant can suggest dinners from your registered fridge and pantry contents. You can update stock by scanning a receipt after shopping, by manually registering items, or by confirming what was used after cooking. When you ask 'what can we have tonight?', the assistant returns three concrete suggestions prioritised by what expires soonest and filtered by your household's saved preferences — not a generic list, but options that are realistic for your specific kitchen and family. Each suggestion includes a simple method and flags any ingredients you're short on.

The connection to Zenframe Meals means the dinner you choose gets added to the weekly plan and generates a targeted shopping list for the few items you need to complete it. The Planner integration tells the assistant whether it's a busy evening — football pickups, a late finish — which pushes simpler, faster suggestions to the top on those nights. Over time the system learns which suggestions your household accepts and ranks future options accordingly, so the quality of the suggestions improves the more you use it.

• Zenframe Assistant suggests dinners prioritised by fridge shelf life and filtered by household preferences
• Chosen dinner adds to the weekly plan and generates a shopping list for only the missing ingredients
• Planner integration pushes faster dinner suggestions on evenings with heavy activity schedules

Practical tips families can start with today

• Confirm fridge stock after each shopping run — two minutes of updates prevents ten minutes of improvisation at 5pm.
• Add a mental 'use by' flag to items with short shelf lives when you put them in the fridge, not when they go soft.
• Ask the AI for suggestions before opening Google or a recipe app — it knows your family's preferences, a search engine doesn't.
• If none of the three suggestions fit, ask for alternatives — the system has enough variety to offer genuinely different options.
• Plan Sunday's dinner last in the week — it's the natural slot to use up whatever the busy week left behind.

FAQ