Veg Box Scheme Planning: Weekly Contents, Swaps and Costs With AI

Plan a veg box by setting a value band for the contents (for example 110 to 120 per cent of the box price at your own prices), listing what will be ready in what quantity, and filling each box from the most plentiful items first. Keep a substitution table for shortfalls and a short swap list customers can use.

The sum that catches new growers out is the gap between price and value. A box that costs the customer $22 and holds $19 of vegetables at your prices feels mean; one holding $30 quietly loses you money. The aim is a consistent, slightly generous box, built from what the field gives you that week, not from a fixed recipe. AI helps with the arithmetic checks, the swap rules and the messages, but the harvest list comes from your eyes and your picker.

Follow me on Instagram@sagnikteaches

Start with the value band, not the vegetables

Most weekly planning goes wrong because it starts with what looks nice in the box. Start with two numbers: the box price and the value band.

Connect on LinkedInSagnik Bhattacharya

The value of a box is the sum of each item's quantity multiplied by the price you would charge for it separately, at your stall or farm shop. Say the box sells for $22 and you want customers to receive 110 to 120 per cent of that: $24.20 to $26.40 of contents. Write the band at the top of the planning sheet. Every box that week gets checked against it before the picking list is printed.

Subscribe on YouTube@sagnikteaches

This is a customer-facing promise even if you never print it. If your price list shows a bunch of carrots at $1.80 and the box sheet uses $1.20, someone will notice. Use the same price list for both, and update it weekly in step with your stall prices. If you run a larger or smaller box, give each its own band, with the small box about two-thirds of the standard.

A main example: building one week's standard box

This is an illustration with invented figures. A small grower sells 120 standard boxes a week at $22 each. This week in midsummer the picker reports what is ready, with quantities and the stall price for each.

The last column, boxes the item can fill, is the quantity divided by the amount per box. The scarce items are the constraints: tomatoes fill only 110 of 120 boxes, and the salad bag only 133 (which is fine). Everything else covers 120 with room to spare.

Now build the box. The planner puts all ten items in. The value is 3.20 + 2.50 + 1.80 + 2.60 + 1.50 + 1.20 + 4.00 + 2.20 + 2.50 + 3.00 = $24.50, which sits inside the $24.20 to $26.40 band. The box would be full, so the weight check matters too: roughly 4 kg including the salad, which fits a standard crate. If the sum had come out at $23, the planner would add an extra bunch of the plentiful item; if $28, take one out and save it for the stall.

The shortfall is tomatoes: 120 boxes need 60 kg and you have 55. That leaves ten boxes without them, and a substitution rule picks up the slack (next sections).

Let AI check the sum and spot the shortfall

The arithmetic belongs in a spreadsheet, with a SUMPRODUCT of quantity and price per box and a comparison with the band. Language models are unreliable at multi-line sums, so do the totals in the sheet and use AI where it is strong: noticing, explaining and drafting. A prompt that works on the planning table:

Here is this week's veg box plan as a table: item, unit per box, stall price, quantity ready, and the number of boxes I need to fill (120). The value band for a standard box is $24.20 to $26.40.
1. List every item where the quantity ready is less than needed for 120 boxes, with the shortfall in boxes.
2. For each shortfall, propose two substitutions from the other items, each within 10 per cent of the original's value, in the same broad category (fruit-veg, leaf, root, alliums), and say what the value would be.
3. Do not change any price or quantity. Do not add items that are not in the table.
Return a short table, not an essay.

[paste table]

A realistic output (illustrative) flags tomatoes (short by 10 boxes) and offers: "Option A: 500 g extra courgettes plus a basil pot, value $2.50 + $1.50 = $4.00. Option B: 400 g green beans, value $4.00." Both land at $4.00, the value of the tomatoes, but look closer. Option A adds a basil pot, which is not in the table at all: the model broke your rule and gave it a $1.50 price of its own. Keep it only if you really have basil (in this example, a few pots bought in and priced at your own stall price). Option B depends on a price per 400 g that does not exist in your table; the model scaled the 300 g price ($3.00 / 300 x 400 = $4.00), which happens to be right, but check it. Always check any item or figure that was not in your input.

Substitution rules written once, used every week

A substitution table is more useful than any weekly improvisation. It says, for each item, what can replace it, at what quantity, and when. Build it once for the season and revise as crops change.

Two principles keep customers happy. First, swap like for like in category and in value, so nobody receives three roots and no greens. Second, never substitute silently: print a note in the box or send an email that names the change. A line such as "No tomatoes this week, the heat ripened the lot early and the picking ran out; here is basil and extra courgettes instead" is better than any apology.

If the harvest changes through the day (a downpour stops the picking), the table lets the packer decide without phoning you. A sign-off rule helps: the packer may use the first swap; the second swap needs a call to the grower.

Customer swaps: a short list and a limit

Customers will ask for no beetroot, double carrots, or "can I have more salad". Saying yes to everything wrecks the packing line; saying no loses customers. The compromise that works for many schemes is a fixed swap list with limits.

• Offer three or four "never send me" items per customer, stored on their record. If a box would include a blocked item, the packer replaces it from the substitution table.
• Allow a small number of weekly swaps, for example two, chosen by a deadline such as midnight on Wednesday for a Friday pick.
• Swaps must be value-neutral. Swapping a $2.50 bunch of beetroot for a $2.20 kale bunch is fine; swapping it for $4.00 tomatoes is not unless the customer pays the difference.
• Keep a paid add-on list (eggs, a flower bunch, a second salad bag) for those who want more.

The sign-up and swap flow is the territory of the farm shares tutorial, which covers form design and pickup reminders. This tutorial stays with what goes in the crate.

The swap-confirmation message

Send a short message after the deadline so customers know what they will get. Here is a prompt and a realistic reply.

Write a short email to a veg box customer confirming this week's box. Facts: first name [first name]; their standing swap: no beetroot (replaced with an extra kale bunch this week); their extra: one dozen eggs, $5.00; box contents: new potatoes 1 kg, courgettes 500 g, carrots, salad bag, cucumber, spring onions, kale, green beans, plus basil pot because tomatoes are short; delivery Friday between 4 and 7 pm. Friendly, under 100 words, no exclamation marks. Sign off [your name].

An illustrative result: "Hello [first name], here is what is in your box this Friday: new potatoes (1 kg), courgettes, carrots, a salad bag, cucumber, spring onions, green beans and a basil pot. We are short of tomatoes this week, so the basil stands in. As usual, kale replaces your beetroot, and your dozen eggs ($5.00) are packed too. Delivery is between 4 and 7 pm. Thanks, [your name]". Check that the list matches the packing sheet exactly. In this draft the standard bunch of kale has vanished from the contents list, folded into the beetroot swap, so a customer who finds two bunches in the crate will wonder why; add it back before sending. A model can drop an item, or add one that sounds likely, without any warning.

From the harvest list to the packing sheet

The weekly planning rhythm for a small scheme can fit in an hour. The rhythm below suits a Friday delivery.

1. Monday: the picker walks the fields and writes expected quantities for the week, item by item. Enter them in the sheet.
2. Tuesday morning: the planner builds the box, checks the value band and the shortfalls, and fills the substitution plan. Ten minutes with a spreadsheet and a short AI prompt.
3. Wednesday midnight: swap deadline. Customer choices are collected.
4. Thursday: update quantities with the actual harvest and fix any new shortfall; produce the packing sheet and the customer messages.
5. Friday: pick, pack and deliver; note any differences between plan and reality.

The "differences between plan and reality" line is the learning loop. If the planned carrots were 80 kg and the picker pulled 68, record it, and over a few weeks you learn how much to trust the Monday estimate for each crop. Crop notes are also useful for the next season; the farm rota and harvest messages tutorial covers sending the picker's list and timings by message.

Waste and the leftover problem

Sometimes the field is generous and you have more of a crop than 120 boxes can use. That is a good problem if you have a plan for it: a larger box for the week, an add-on offer ("extra courgettes, $4 a kilo this week only"), the stall, or a restaurant order. If you have none of these, a surplus becomes waste. A short prompt helps you generate options:

I have 40 kg more courgettes than my 120 boxes need this week, and they will not keep beyond five days. List six ways to move them within five days at a small veg business (customer offers, local buyers, preserving, boxes), with a one-line note on effort and likely return for each. Do not invent prices.

The answer is a brainstorm, not a plan, so pick one and test it. Over a season you will see the pattern: a glut of courgettes in July, of kale in autumn. Use it to plan standing offers rather than react. The ideas in the food-waste forecasting tutorial carry over to a produce business.

A quick weekly cost check

The box can hit its value band and still lose money, so check the sums once a week. Here is an illustrative check for the 120-box week at $22 a box.

From the $2,218 left you still have to pay for growing the crops: seed, water, field labour, machinery and your own time. If growing costs run at $1,800 for that week's harvest, the margin is about $418, or 16 per cent of revenue. A different customer mix, one more bunch of high-value tomatoes per box, or a 10 per cent drop in labour hours changes it quickly, which is why the weekly check matters more than the annual one. For a deeper model of price and margin on a subscription-type product, the subscription box margin model tutorial goes further, and the sums should stay in a spreadsheet.

Keeping customers once the novelty fades

Box schemes lose customers to boredom as much as price. Three signs show up in your data: customers skip weeks more often, swap lists grow, and complaints mention "too much kale" or "again beetroot". You can see these patterns from the swap records and messages if you keep them.

Paste a month of anonymised swap requests and comments into a chat assistant and ask for the themes: which items are blocked most, which comments repeat. If a quarter of customers block beetroot, the standard box should carry less of it, with beetroot moved to an add-on. If many comment on the "samey" box in late autumn, put a recipe card in for the three heaviest items. When a customer cancels, ask why in one line; the tutorial on cancel reasons and save offers shows how to turn those answers into a small offer, such as a pause rather than a cancel.

An organic grower's extra step

If your produce is certified organic and boxes carry that claim, each lot in a box must stay traceable. Add a lot or field code per item on the packing sheet and keep the sheet. The organic records tutorial shows a trail from sale back to field. Where you buy in to fill a shortfall, such as the tomato substitution, the substitute must meet the same standard if the box is sold as organic, so a bought-in basil pot needs its own supplier proof or must be left out of an organic claim. Do not ask AI whether a substitute qualifies; ask your certifier.

When the box over-delivers by accident: a plan that drifted

A pattern that shows up in many small schemes: the box slowly grows. In spring the planner adds a bunch of radishes because they look good. In June a second salad bag goes in because lettuce is bolting. By August nobody has re-checked the value band, and the box is holding $29 of produce at stall prices for a $22 price. Nobody complained, so nobody noticed, and, if growing costs rise in step with the extra produce, the weekly margin slides from about 16 per cent to about 3 (roughly $88 on $2,640 of boxes).

The fix is dull and works: a weekly line on the planning sheet that shows the box value against the band in red or green, and a rule that any addition above the band needs a removal. A formula in the sheet does the checking; the AI prompt above is only for the shortfall. Look at the last eight weeks of totals and you will see the drift before it costs you a season.

Small box, large box and the winter version

Many schemes run two sizes, and the plan for the second is not a halving of the first. A small box at $14 might have a band of $15.40 to $16.80. Use the same ten items but fewer of each, then remove awkward items, such as a single tomato punnet that cannot be halved, and replace them with items that scale, such as a smaller kale bunch. Check that each unit is realistic to pack: half a cucumber is not.

Winter changes the problem entirely. Fewer crops are ready, storage vegetables dominate (potatoes, carrots, beetroot, onions, squash, cabbage), and value per kilo is lower, so the box gets heavier for the same price. A simple winter rule is to fix a core of three storage crops that appear every week and rotate two or three seasonal extras. Leaves and herbs from the polytunnel then carry the value band. Ask AI to suggest recipe cards from the core items, one per week, and have someone who cooks read them, since a plausible-sounding method that does not work will cost you a customer faster than a missing tomato.

A plan for the week the picker is ill

The estimate on Monday comes from one person's walk through the field. If that person is away, write the plan from last year's notes for the same week and halve your confidence in it. Keep one fallback box, built from the three or four items that almost never fail (potatoes, carrots, onions, kale), and publish it as the default for the week. Customers get a simple box and a short, honest note, and you avoid last-minute guesses that end in short boxes.

Checking the scheme is healthy

After a month of using the plan, look at five figures. How often did a box fall outside the value band, and in which direction? How many boxes needed a substitution, and did customers mention it? What share of picked produce ended as waste or low-price stall sales? How many customers used their swaps, and which items did they block? And what was the weekly margin after direct costs?

The answers tell you where to spend effort. A box that is often over-value means you are giving away money; often under-value means customers will leave. High substitution rates point at a crop plan problem rather than a packing one. The planning routine here takes about an hour a week once the sheet is built, and each week's notes make the next season's sowing plan a little less guesswork.

Questions from growers running a veg box

How much value should a box hold compared with its price?

Many growers aim for contents that would cost somewhat more than the box price at their own shop or stall prices, so customers feel they get a fair deal. Choose a band, such as 110 to 120 per cent of the price, and check each week's box against it.

Should customers choose their own items?

Full choice makes packing slow and wastes the best of the harvest. A good middle is a set box with a short list of items each customer can swap out, and a limit on how many swaps are allowed each week. Offer a paid add-on shop for extras.

What if a crop fails the week before?

Keep a substitution table with like-for-like swaps in the same category and value, and tell customers the rule when they sign up. Send a short note in the box or by email naming what changed and why. Do not swap silently.

Can AI plan the box from my harvest list?

It can propose a mix and explain the swaps, but it should work from your real quantities and prices, and you should check totals in a spreadsheet. It cannot see what is actually ready in the field, so the picker's list stays the starting point.

How do I handle allergies and dislikes?

Record them per customer in a field the packer sees on the packing sheet. Treat true allergies as a safety matter and handle them by hand, not by a swap rule alone. Ask a food-safety adviser about labelling and cross-contact in your packing area.

Further reads

• Subscription Box Curation: Match Items to Surveys With AI — Match items to what each customer says they like.
• AI for Farm Shops: Weekly Boxes, Seasonal Labels and Pre-Orders — Weekly boxes and seasonal labels in a farm shop.
• Is AI Worth It for a Small Subscription Box? Five Jobs First — Five jobs to hand to AI first in a box business.
• AI in Food Retail: Jobs a Deli or Farm Shop Can Hand Over First — Other jobs a farm shop can hand over first.
• Can Farmers Use AI for Farm Records and Paperwork? — Farm records and paperwork AI can help with.
• Farm Accounting Software: QuickBooks, Xero or a Farm App With AI? — Where box income and farm costs should be recorded.
• AI for Farmers Market Vendors: Pre-Orders, Pricing and Weekly Posts — A weekly cycle for a market stall: pre-order sheet, price list by pack size and a Saturday post template, with AI doing the drafting.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: general produce-box practice; no vendor pricing is quoted in this tutorial. Check food-safety and labelling rules for your trade with your local authority.