AI for Farmers Market Vendors: Pre-Orders, Pricing and Weekly Posts

A farmers market vendor can use AI for three weekly jobs: turning pre-orders into a pick list and total, working out what each pack size earns from your costs, and drafting the Saturday post from a short template. Give it your numbers, check its sums, and keep customer names out of it.

None of this needs a special app. A spreadsheet, a free assistant and about 45 minutes midweek will do it. What matters is the order of work: the pre-order sheet first, because it tells you how much to pick or bake, then the price list, because pack sizes drive margin, then the post, because it should announce what the first two have decided.

Follow me on Instagram@sagnikteaches

The weekly cycle on one page

Here is the shape of a week for a stall that sells vegetables, eggs and jam. Substitute your own produce. The cycle is the point, not the products.

Connect on LinkedInSagnik Bhattacharya

The Saturday leftovers note is the quiet engine. After a month it tells you, item by item, how much extra to bring beyond the pre-orders. That makes the next job, planning quantities, evidence-based; if you want to take that further, farmers market stock planning with AI covers it in depth.

Subscribe on YouTube@sagnikteaches

Step 1: a pre-order sheet that totals itself

Pre-orders arrive in every shape: texts, direct messages, a comment under a post, a note passed at the stall. The job is to get them into one table. Keep that table in your own spreadsheet with these columns: order number, customer first name or initials, contact route, item, quantity, pack size, paid or not, collected or not.

Use AI for the typing, not the storage. Paste in the raw messages with names replaced by the order number, and ask for rows.

Turn these customer messages into table rows. Columns: order number, item, pack size, quantity. Only use items from my availability list below. If a message asks for something not on the list, or the quantity is unclear, put the order number in a separate list called QUERIES with one line saying what is unclear. Do not guess.

AVAILABILITY LIST:
Eggs, half dozen; Eggs, dozen; Salad bag 150 g; Carrots, 1 kg bag; Strawberry jam 227 g jar; Chutney 227 g jar

MESSAGES:
#14: two dozen eggs and a jam please
#15: any carrots left? 2 bags. also a salad
#16: usual please
#17: 6 eggs x3, chutney

Illustrative output: rows for #14 (eggs dozen x2, jam x1, but jam flavour not stated, so also listed in QUERIES), #15 (carrots 1 kg x2, salad bag x1), #17 (eggs half dozen x3, chutney x1), and a QUERIES list saying "#14: which jam?" and "#16: what is 'the usual'? Not in my records." The useful part is the QUERIES list. The assistant could not know what #16's usual is and, because you told it not to guess, it said so. Without that instruction many assistants would invent a usual.

When the sheet closes, ask it to total by item and compare with what you expect to have. Then check the totals yourself with a SUM in the spreadsheet. A model can slip on addition across a long list, and a wrong total here means a wrong crate on Friday.

Thursday evening: closing 22 pre-orders

Take an illustrative stall with 22 pre-orders closed on Thursday night. The sheet shows these totals against what is available to harvest or make:

The sheet exposed a shortfall of two salad bags that you would otherwise find out about on Saturday morning. A pick list that says "salad: 18, picked so far 16" gets you to the row and to a decision: pick more on Friday, or contact two customers and offer a swap. A short message for that situation, written once and reused, saves worry:

Write a short, friendly message to a customer whose pre-ordered item is short this week. Item: salad bag. They ordered 1. I can offer a bigger bunch of herbs or a refund of the amount paid, their choice. No apology paragraph, no emojis, under 40 words. Sign off with [your name].

Step 2: a price list by pack size

Most stalls set prices by feel and by what the next stall charges. That is not wrong, but it hides two traps. A smaller pack can earn less per hour of your labour even when the per-gram price looks higher, and the pack that customers choose is often the one that gets you less margin. A table of each pack with its cost and its earnings makes both visible.

Build the table from facts only you have: ingredient or growing cost per unit, packaging cost per unit, the time to pack, and the price you charge now. Then ask the assistant to do the arithmetic and lay out the options. Always say that the figures are yours and it must not add any.

I sell strawberry jam. Here are my costs and prices. Do the arithmetic only; do not add any figures of your own. For each pack size show: total cost per pack, price, profit per pack, margin as a percentage of price, and profit per pack if I discount by 10 per cent for buying three. Then tell me which pack earns the least per pack and which has the lowest margin.

114 g jar: fruit and sugar 0.55, jar and lid 0.38, label 0.07, price 3.50
227 g jar: fruit and sugar 1.05, jar and lid 0.46, label 0.07, price 5.50
340 g jar: fruit and sugar 1.55, jar and lid 0.58, label 0.07, price 7.50

Doing the sums by hand: the 114 g jar costs 1.00 and sells at 3.50, a profit of 2.50 or about 71 per cent of the price. The 227 g jar costs 1.58 and sells at 5.50, a profit of 3.92 or about 71 per cent. The 340 g jar costs 2.20 and sells at 7.50, a profit of 5.30 or about 71 per cent. Margins match, so the price ladder is consistent. A multi-buy at 10 per cent off drops the 227 g jar's profit to 3.37 (price 4.95, cost 1.58), still about 68 per cent. This is the kind of check a stallholder can do in a minute and then decide a "three for 15" offer is safe.

The sum above ignores your time and the stall fee, which many stalls leave out and later regret. Add a column for packing time and ask for profit per minute. If a 114 g jar takes as long to fill, lid and label as a 340 g jar, its profit per minute is less than half. That is a sound reason to keep the small size as a gift or sampling size and not push it. For the costing method behind this, the batch, yield and margin sheet goes into it properly.

Rounding prices people can say and carry change for

A price of 5.43 is correct and useless at a stall. Ask the assistant for the nearest whole or half-unit price both above and below the calculated one, and what each does to the margin. Then choose the one you can make change for without a calculator. Shoppers also compare, so check it against two nearby stalls by looking, not by asking the assistant, which has no knowledge of your market.

Step 3: the Saturday post template

A weekly post fails in one of two ways: it is skipped, or it is bland. A template solves the first and your facts solve the second. The template has five slots, and AI fills only the wording around facts you give it.

1. What is new or best this week, one line, from you.
2. What to pre-order and by when, with the deadline written plainly.
3. One price or deal, from your price list.
4. One true detail about how it was grown or made this week.
5. Where to find the stall, written once and reused.

Write a short Saturday post for a market stall, under 90 words, warm and plain, no emojis, no hashtags beyond one. Use only the facts below. Do not invent weather, awards, quantities or claims about health or freshness beyond what I state.

NEW THIS WEEK: first sweetcorn of the year
PRE-ORDER: reply by Thursday 8 pm for Saturday collection at the stall
DEAL: three jars of jam for 15
DETAIL: the corn was picked on Friday evening after the rain
STALL: usual spot by the east gate, 8 am to 1 pm

Illustrative output: "Sweetcorn is in this Saturday, picked on Friday evening after the rain. Pre-order by Thursday 8 pm and we will have it bagged with your name on it. Jam is three jars for 15 while stocks last. Find us by the east gate, 8 am to 1 pm." Usable as is, but two edits are common: "bagged with your name on it" is a promise only if you do that, and "while stocks last" is padding you can cut. Read the draft for anything the assistant added that you would not say at the stall.

Save the post you like best as a model. Next week, paste it and ask for a new post in the same voice with this week's facts. That keeps the style steady without a long instruction. If you want a week of posts and graphics together, the Canva weekly batch tutorial shows the design side, and the pre-publish checklist for AI-drafted posts is a good last look.

More stall situations, worked through

A regular who always orders the same

Message #16 above, "the usual", is where a small memory sheet pays for itself. Keep a tab with each regular's initials and standing order, such as "R.: dozen eggs, one salad". When a message says "the usual", you look it up in seconds. Do not feed the whole customer list to the assistant to remember; a plain lookup does that job without sharing anything.

Before and after: a pre-order announcement

Before: "Pre orders open for the weekend message me what u want." After: "Pre-orders are open until Thursday 8 pm. Message the item, the pack size and how many. Eggs, salad bags, carrots and jam are available. I will confirm each order with a number." The second version asks for exactly the three things that the sheet needs, which cuts the back-and-forth by about a message per customer. Ask the assistant to rewrite your announcement so that it requests item, pack size and quantity, then check that it kept your deadline.

A bakery stall working to a cut-off

A baker whose bread is baked to order has a harder version: the oven schedule depends on the count. The pre-order total goes straight to a bake plan, loaves by type, with the dough batch size as the unit. Say a batch makes 12 loaves and pre-orders for sourdough total 19. Two batches make 24, leaving five for the stall, which you can compare with the leftovers note. Ask the assistant to round totals up to whole batches and show the stall surplus, then check its batch arithmetic.

Prices after a poor season

If a cold spring halves the strawberry crop, your cost per jar rises. Re-run the pack-size table with the new fruit cost and compare. A rise from 1.05 to 1.60 in fruit and sugar for a 227 g jar takes the profit from 3.92 to 3.37 at the same price and the margin to about 61 per cent. You can decide to hold the price, add 50 pence, or retire the small jar for the season. The table makes it a decision with numbers instead of a worry.

A mistake in a total that cost a customer a crate

One illustrative stallholder pasted 40 orders into an assistant and asked for totals. The assistant reported 31 salad bags. The real count was 36, because five orders said "salad x2" and the model counted each line as one. The stall was five bags short on the day. After that the stallholder added a rule: totals come from a spreadsheet formula, and the assistant is only used to put messages in rows. A quick way to catch this kind of slip is to ask for the row count and the total side by side, and to test one item by hand every week.

Three weeks of a stall that starts from nothing

If you have never run a pre-order sheet, do not build everything at once. An illustrative ramp for a stall selling eggs, vegetables and preserves looks like this.

Week one: only the sheet. Announce pre-orders for three items you are sure of. Collect orders by message, paste them into the assistant with the availability list, and total them yourself in the spreadsheet. The aim is to learn how customers phrase orders, so read the QUERIES list carefully and note which questions come up twice. Perhaps half your customers will not say a pack size. Add the pack size to the announcement wording and the problem shrinks the following week.

Week two: add the price table. Build the pack-size table for your three busiest products only. Look for the pack with the lowest profit per minute and either raise it, bundle it, or stop making it. Resist doing every product; the busiest three usually carry most of the revenue.

Week three: add the post. Use the five-slot template and keep the best result as a model. By now you have real facts for the slots: the pre-order count, the deadline that worked, the deal you tested.

Thirty minutes of review after week three is worth having. Ask what you would drop if you had to cut one step. Many stallholders find the post is the least valuable of the three, because pre-order customers already know what is coming, and the sheet and price table are the pieces that change money. Keep what earns its time and drop the rest.

When the assistant is the wrong tool

Some jobs are better done without it. Stock on the day, at the stall, is a pencil job. Anything involving a customer's allergy is yours. Anything that depends on a local market rule, a licence or a hygiene requirement is a question for the market organiser or the authority that licenses food sellers, and an assistant's answer should not substitute. And any figure that goes on a sign, a label or a receipt gets checked by you against the real costs. Treat the assistant as quick, tireless and prone to confident slips, and give it the jobs where a slip is cheap to catch.

Keeping customer details out of the tools

You need only items and quantities to build a pick list. Names, phone numbers and addresses are not needed by the assistant. Replace each with an order number before pasting, and keep the key in your own sheet. If you use a form for pre-orders, look at where the responses are stored and who can see them, and what the assistant tool does with what you paste: ChatGPT's Free, Plus and Pro plans have an "Improve the model for everyone" setting that is on by default and can be switched off, and Claude's consumer plans have a similar model-improvement switch in their privacy settings. If you collect payments or hold allergy information about customers, ask a qualified adviser what rules apply to you before storing it anywhere.

Allergies deserve a special mention. If a customer states an allergy in a pre-order, that message is for you to handle by hand with the ingredient list in front of you. Do not ask an assistant whether a product is safe for them. It can draft the question you ask the customer, but the answer comes from your recipe and your handling of ingredients.

Three numbers that show the stall routine pays

After four weeks, three numbers should be visible in your sheet. The first is the share of sales that came from pre-orders, which should rise as customers learn the habit. The second is the leftover quantity per item, which should fall as you plan against orders. The third is the time spent on admin, which should settle near 45 minutes a week. If leftovers are not falling, your stall margin on top of the pre-orders is too high; trim it item by item, not across the board.

If pre-orders stay low, the fault is usually the announcement, not the tool. Make the deadline clearer, shorten the message needed to order, and offer a small reason to pre-order, such as a guaranteed pack of the item that sells out first. If you later sell a regular box rather than loose items, veg box planning picks up where this leaves off, and weekly boxes and labels for a farm shop shows a related set-up.

Questions from stallholders about using AI each week

Which AI tool is enough for a market stall?

A free or entry-level plan of a general assistant such as ChatGPT, Claude or Gemini, plus a spreadsheet you already use. The work is drafting text and tidying a table once a week, which needs no special app. Pay only if you hit usage limits.

Should I paste customer names into an AI tool?

Avoid it. The assistant only needs items and quantities to total an order sheet. Keep names and phone numbers in your own sheet and use order numbers or initials in anything you paste. Check the tool's data settings before you paste anything personal.

Will AI set my prices for me?

It can do the sums and show options, but it cannot know what your neighbours charge or what your regulars accept. You supply costs and the price you want to land near; it shows you what each pack size earns. The decision stays yours.

How do I stop the Saturday post sounding generic?

Give it three things it cannot invent: what is actually in the crate this week, one thing that changed (a late frost, a new variety) and the pre-order deadline. Then cut every line that would fit any stall in any week.

What if pre-orders do not sell out the stock?

That is the point of the sheet: it shows you the committed amount before you pick or bake. Take pre-orders plus a stall margin, say 15 to 25 per cent extra, and record what is left each week so the margin is based on your own sales.

Further reads

• AI for a Cottage Food Business: Labels, Pricing and Order Forms — Labels, pricing and order forms for home-produced food.
• Selling Beef Boxes Direct: Hanging Weight, Cut Sheets and Pricing With AI — Price direct-sale meat boxes by weight.
• Should a Small Business Use AI to Write Its Social Media Posts? — Decide how much of your social posting to hand to AI.
• How to Use AI in a Bakery Business: Wholesale and Standing Orders — Standing orders for a baker selling beyond the stall.
• Subscription Box Pricing: Build a Margin Model With AI — Model margins if you sell a regular box.
• AI for Butchers: Christmas Pre-Orders, Cut Lists and Pick-Up Slots — Set weight bands, cap pick-up slots, turn an order export into a cut list and send confirmations with AI, while allergen and label checks stay with you.
• Starting a Chocolate Business: Costing, Labels and Where to Sell With AI — Per-bar costing, allergen label checks and a margin comparison of farmers market, online and wholesale, with illustrative numbers.
• Street Food Stall Business Plan: Pitch Fees, Margins and Stock With AI — Cost every portion, compare pitch fees on wet, normal and busy days, and plan stock and gas for a street food stall, with illustrative numbers and AI prompts.
• Food Truck Ordering App: Pre-Orders, Pitches and Where AI Helps — Compare Square, Toast and food-truck platforms on pre-order slots, location sharing and fees, then see where AI helps and what an app will never do.
• Microgreens Business Plan: Tray Costs, Restaurant Orders and Waste With AI — A microgreens plan built from one tray: what it costs, what it yields, what a standing restaurant order is worth and how a waste log changes the sums.
• Pre-Order Policy Checklist: Draft Yours and Check Delivery Rules — A checklist for an online store's pre-order policy, a prompt that drafts it, and the 30-day delivery rules to check it against.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: stallholder workflow described in general terms; training settings from OpenAI's help article on data controls in ChatGPT and Anthropic's privacy centre, checked November 2026. No figures in the worked examples are real; they are illustrations.