ChatGPT for Screen Printing Shops: Quotes, Art Prep and Customer Emails

A screen printing shop can give ChatGPT its price structure and art checklist, then use it to draft quote explanations, art-prep notes and customer emails from the details of each job. The figure itself should come from your price sheet or shop software, and a person should check every draft before it leaves.

The trap is asking the tool to work out the price. A model will happily produce a number that looks right and is wrong by a screen charge or a quantity break. Use it for words and structure, and keep the arithmetic somewhere that can be audited. The sections below show that split in practice for a shop with a small team and a steady flow of club, school and business orders.

Follow me on Instagram@sagnikteaches

The shop used in the examples

Take an illustrative shop with three staff: an owner who quotes, a press operator and a part-time person who handles art and emails. It prints on cotton tees and hoodies using a manual or semi-automatic carousel, quotes thirty or so jobs a week and wins roughly a third of them. Every quote needs the same questions answered: how many garments, which garments, how many colours, where on the garment, whether the artwork is ready and when it is needed.

Connect on LinkedInSagnik Bhattacharya

Time disappears in two places. The first is the back-and-forth to collect those answers. The second is explaining the price to someone who thinks a printed tee should cost what a blank one does. Both are text jobs, which is where ChatGPT helps most.

Subscribe on YouTube@sagnikteaches

Write the pricing logic down first

Screen printing pricing follows a pattern most shops share: a setup charge for each screen (one screen per colour per print location), a per-piece print price that falls as the quantity rises, and extras for dark garments needing an underbase, oversized prints, specialist inks, labels or folding. Before ChatGPT sees anything, write your version as plain rules. Here is an illustrative sheet with invented numbers; use your own.

The structure matters more than the numbers. Notice how the rules use colour count as the driver, because colour count drives screens, ink and press time. Dark garments quietly add a colour for the underbase, and a surprising number of quotes miss that, which is why a prompt should ask about garment colour explicitly.

A quote prompt that asks before it guesses

Put the price sheet in a ChatGPT Project so you are not pasting it each time. OpenAI's help centre states that project sharing works on every plan, and file limits vary by plan, so a one-page rules file fits. Then give the Project short instructions.

You help a small screen printing shop prepare quotes.
Use only the price sheet file. Never do the final arithmetic yourself:
instead, list the line items and the numbers to add, and show a
calculation I can paste into a spreadsheet.
If the customer's message lacks any of these, ask for them in a short
list rather than guessing: quantity, garment type and colour,
number of print locations, number of ink colours per location,
artwork status, deadline.
Flag any dark garment as needing an underbase colour.

Now paste a typical enquiry, with personal details removed:

Hi, we need about 60 hoodies for our running club, black,
our logo on the chest and the club name on the back. Need them in
three weeks. Logo is red and white.

An illustrative response lists what is known (60 pieces, hoodies, black, two locations, three weeks) and asks what is missing: how many colours in the back print, whether the logo file is vector, and what hoodie style and sizes. It adds a note: "Black garment: red and white chest print needs underbase, so treat as 3 colours on chest." That is the behaviour you want, a model that notices the dark garment.

The calculation it produces for the spreadsheet might read: chest, 3 screens at $20 equals $60 setup; back, 1 colour white plus underbase equals 2 screens, $40 setup. Print price at 48 to 143 pieces: $2.40 for one colour plus $0.60 for each extra colour. Chest at 3 colours is $2.40 plus $1.20, which is $3.60. Back at 2 colours is $2.40 plus $0.60, which is $3.00. Per piece, $6.60, times 60 is $396, plus $100 setup, plus garments. Always rerun it in the spreadsheet. If the figures disagree, the spreadsheet is the version you trust, and the disagreement tells you the prompt needs a tighter rule.

Turn the number into a quote the customer can follow

Once you have a confirmed figure, ask for the customer-facing message. The goal is a quote that explains itself, so the buyer does not need a phone call to understand why the price is not just the blank hoodie cost plus a logo.

Write a short quote email using these confirmed numbers:
60 black hoodies, chest print 3 colours (red, white, plus underbase),
back print 2 colours (white plus underbase), setup $100,
total $[X], delivery in 15 working days from artwork approval.
Explain in two sentences why a dark garment adds an underbase.
Ask them to confirm sizes and approve the artwork proof.
Tone: friendly, plain. 150 words maximum. Sign as [your name].

The reply usually needs light edits: a stray "delighted to help" removed, the delivery date confirmed against the press schedule, and the total double-checked. Do not let it add promises such as "free proof included" unless that is shop policy.

Art-prep: a checklist the model can apply to what you tell it

Art problems cost more than any other stage: a low-resolution logo discovered after screens are burnt wastes screens, ink and a day. A checklist catches it early. Write yours as questions, and let ChatGPT turn the customer's answers into a note for your art person.

1. File type: vector (AI, EPS, PDF, SVG) or raster (PNG, JPG)?
2. If raster, the pixel size at print size; as a rule of thumb, aim for 300 pixels per inch at the actual print dimensions, and ask for the original where you can.
3. Number of ink colours, and whether any gradients, halftones or photos are included.
4. Fonts converted to outlines?
5. Print size and position (width in inches from the collar for a chest print).
6. Garment colour, which sets the underbase and whether white or light inks are needed.
7. Pantone references if brand colour matching matters.
8. Minimum line thickness and any tiny text that may close up on the screen.

The prompt:

Here are a customer's answers to our art checklist: [paste answers].
Write an internal art-prep note listing: what is ready, what needs
fixing before screens are made, the questions to send back to the
customer, and the risk if we proceed as is. Do not assume the file
is suitable if resolution or vector status is unknown.

A typical output: "Ready: colours, placement. Needs fixing: logo supplied as a 400-pixel PNG, too small for a 10-inch chest print. Ask for vector or a larger original. Risk: soft edges and fuzzy text if we enlarge it." That is a note your art person can act on in seconds, and a customer email that politely asks for a better file follows from the same list.

No customer artwork on consumer plans

Customer logos belong to the customer, and some are confidential, such as an unreleased product or a club's new branding. Keep image files out of personal ChatGPT accounts. On personal Free, Plus and Pro accounts, OpenAI says data sharing for model training is on by default; you can switch off "Improve the model for everyone" under Data controls, but treat that as a mitigation and not a licence to upload. ChatGPT Business does not train on your business data by default, according to OpenAI's business pricing page, which lists a Standard seat at $20 a user a month billed annually or $25 billed monthly, for teams of 2 to 200. Check the page before budgeting.

A workable rule for a small shop: describe artwork in words and checklist answers, never the file; never paste a customer's name with their order details if the name is not needed; and keep the shop's own price sheet, which is not sensitive to the customer, in the Project. If you do want the tool to look at artwork, use a business workspace and tell the customer in your terms that you may use AI-assisted tools for admin.

Reorder emails: the cheapest win in the shop

Repeat customers are the best margin. A club that ordered kit last autumn will probably order again, but nobody remembers to ask. Keep a simple sheet of past orders with date, quantity, garment, and artwork reference, then ask ChatGPT to draft a reminder from each row.

Write a reorder email to [first name] at [organisation type].
Last order: 60 black hoodies, chest and back print, delivered
in September. Offer: same artwork, no new setup charge if ordered
by [date]. Mention we need 3 weeks for production. Ask whether
sizes or quantity have changed. Two short paragraphs, friendly.

Before and after, in a sentence. Without the system, reorders arrive when customers remember, often with a one-week deadline and a rush fee argument. With a reminder sent six weeks before the usual event, a reasonable share of customers order early, and the shop smooths its press schedule. The sheet takes an hour to set up and ten minutes a week to run.

Check each draft against the order record. The common mistake is a wrong quantity: a reorder email quoting 60 when the last order was 48 triggers a dispute. If you waive the setup fee, make sure the artwork really is unchanged and the screens are still on file or reclaimable.

Other emails that eat the day

Beyond quotes and reorders, four message types repeat constantly, and each is worth a saved prompt in the Project.

• Proof approval. "Please check spelling, colours and placement; we print exactly what is approved." Ask the model to produce a proof email that lists the specific checks and the deadline for sign-off.
• Delay notice. When a garment delivery slips, a calm message with the new date and options beats silence. Give the model the facts and tell it not to apologise more than once.
• Quote follow-up. Three days after the quote, a short prompt: "Does the price work? Happy to adjust the quantity to reach the next price break." The break-point suggestion is the useful part, and your price sheet supplies it.
• Collection or dispatch. Tracking details, how many cartons, what to check on arrival and how to report a fault within a set number of days.

For the follow-up in particular, see following up silent quotes with AI, which covers timing and tone in more detail.

A second quote walked through: one colour on light tees

Not every job has a dark garment. Say a school parents' group asks for 150 white tees with a one-colour navy logo on the chest. Setup is one screen at $20. At 144 and over, the one-colour print price is $1.60, so 150 pieces cost $240 to print. Add $20 setup for $260, plus garments. With no underbase and no second location, the sum is short, and that is useful for testing: if the model flags an underbase here, your instructions are too eager and need a line saying "light garments need no underbase".

Now see what a price break teaches. If the group asked for 140 tees, the print price would be $2.40 each, giving $336 plus $20 setup, $356. Ordering ten more tees drops the bill by $96. That is a good moment for the follow-up prompt: ask the model to draft a note pointing out the break, using the numbers you supply. Customers rarely mind a shop that saves them money, and the shop still sells ten extra garments.

A difficult email: the customer who changes the art late

Late changes are where small shops lose margin. A customer approves a proof, screens are burnt, then asks for a different shade of red. Ask the model for a reply that states the position without heat. Give it the facts: the proof was approved on a date, the screens exist, a colour change means remaking one screen at $20 and may move the delivery by two days.

Write a calm reply to a customer who wants to change a colour after
proof approval. Facts: proof approved [date], screens already made,
a change costs $20 for one new screen and may add 2 working days.
Offer two options: keep as approved, or change and accept the cost
and delay. Do not apologise more than once and do not blame them.

What comes back is usually good, though I would change any sentence that implies fault. Keep a copy of your approval wording in the Project, so that "we print exactly what is approved" is policy and not a surprise.

A realistic mistake and how it showed up

Picture a shop that pasted an enquiry containing the line "about 50 shirts" and let the model assume 50. The quote used the 48 to 143 break at $2.40. The customer actually ordered 44 and the shop honoured the quote, losing $1.10 a piece on the print price, since 24 to 47 pieces costs $3.50 by its own sheet: about $48 across the job. Nobody noticed until the end-of-month review, where the margin on that job was visibly thin.

The cure is not a smarter model. It is the line in the instructions that says "if quantity is approximate, ask for an exact number or state the break the quote depends on". Add a sentence to the quote email saying the price holds for the stated quantity only. Small guard rails of that kind do more than longer prompts.

Where the arithmetic goes wrong, and a second check

Models are fluent at wording and fallible at multi-step sums, which is why the quote workflow separates them. The usual slips are: missing the underbase on a dark garment, applying the wrong quantity break (48 to 143 versus 144 and over), charging setup twice for the same colour and location, and forgetting the rush surcharge. A second check takes thirty seconds: recompute from the price sheet in a spreadsheet, and compare. For a deeper explanation of why this happens, why AI is bad at maths and what to check in quotes is worth a read.

A realistic test before you trust the setup: run ten past quotes through the Project and compare its line items with what you actually charged. Count mismatches. If three or more differ, fix the price sheet rules, because ambiguity in your sheet is usually the cause, not the model. Re-test until the mismatches come from genuine judgement calls, such as a favoured customer discount, which belong with the owner.

What a week looks like with this in place

Return to the three-person shop. Monday: enquiries land in the shared inbox, the art person pastes each into the Project, and the model returns a missing-information list and a draft reply. Tuesday: the owner reviews drafts, runs the numbers in the spreadsheet and sends quotes. Wednesday and Thursday: art-prep notes feed the screen-burning schedule. Friday: reorder reminders go out for jobs that are six weeks old. The owner estimates the time saved on typing and chasing; the shop does not need to claim a figure, but can measure it by timing ten quotes before and after.

The same approach suits similar shops. An embroidery business replaces colours with stitch counts, and a sign maker uses area and material. If you want a broader view, see whether a print shop can use AI to price jobs and check artwork, and for converting spoken notes into a job sheet, turning on-site voice notes into job sheets and quotes. The wider picture of AI in a small firm is in the AI guide.

Setting up the Project in twenty minutes

Create a Project named "Quotes and emails". Add the price sheet and one page of shop rules (payment terms, deposit, turnaround, what is and is not included). In the Project instructions, paste the short prompt above. Then save five reply templates as files: quote, proof, delay, follow-up, reorder, each with a sample customer name written as [first name].

Do not use a Custom GPT for this. OpenAI's retirement FAQ says custom GPTs are scheduled to retire on 11 December 2026, with creators moved to plugins, so a Project is the lasting home. Review the Project's files each quarter when supplier prices or your rules change, and put a "last updated" date in the price sheet's first line, so nobody quotes from an old version.

Finally, agree who is allowed to send. A sensible rule: the model drafts, the art person edits, the owner approves anything over a set value, say $500, and every sent quote is logged with its spreadsheet. That small gate is the difference between a time saver and a source of expensive misquotes.

Screen printing and ChatGPT questions

Can ChatGPT calculate a screen printing quote?

It can follow a price sheet you give it, but treat the arithmetic as a draft. Large language models make slips with multi-step sums, so keep the real price in a spreadsheet or your shop software and use ChatGPT to write the explanation and the follow-up around a number you have already checked.

Should I upload customer artwork to ChatGPT?

Not on a personal or free plan. Customer logos can be confidential or owned by someone else. Describe the art in words, or paste your checklist answers, and keep the files in your own systems. If you want the tool to look at images, use a business workspace and tell the customer.

What should I tell ChatGPT about my pricing?

Give it the structure, not the whole business: quantity breaks, price per colour, setup per screen and extras such as dark garments. Mark any number you are unsure of as 'check', so it asks instead of inventing a figure.

Can it write the reorder emails automatically?

It drafts them well when you give it the last order details. A person should still press send, because a reorder email quoting the wrong quantity or a changed price causes a dispute. Keep one template per situation and review each draft before it goes.

Further reads

• How to Design Flyers and Printed Marketing With AI — Design help for the marketing side of a print business.
• How a Plumber Can Save Five Hours a Week on Admin With AI — A comparable small-trade admin routine.
• The Intake Questions That Make AI Cleaning Quotes Accurate — Intake questions that make a quote accurate.
• How Removal Firms Turn Customer Videos Into Quotes With AI — Another trade turning customer material into quotes.
• AI for Interior Design Studio Admin: Sourcing, Quotes, Updates — Quotes and updates for a small studio.
• AI for Printing Companies: Proof Approvals and Job Status Updates — Proof email, approval log, reminder and ready-for-collection message: the message flows a print shop can draft with AI, and the checks that keep them accurate.
• AI in Sports Business: Team Kit Orders With Sizes, Names and Numbers — One club, 30 players, three squads: an order form, a duplicate-number check, a supplier order sheet and parent reminders, with sums and sample prompts.
• AI Print Estimating Software Compared: What a Print Shop Pays — AI estimating versus shop-management software for a small print shop, a monthly cost stack, and a ten-job accuracy test.
• Print Production Scheduling Software: Where AI Actually Helps — Seven print software vendors read for AI claims, a capacity model for presses and finishing, a due-date sum, list prices and a one-week board trial.
• Web-to-Print Software: Online Storefronts, Instant Quotes and AI — Storefront plans from $49 to $439 a month, how instant-quote rules work, and where AI genuinely helps with quotes and proofs.
• AI for Printing Business Repeat Orders: Reminders and Re-Quote Emails — Turn your order history into reminder emails and re-quotes: find who is due, draft the message with AI, check the price, then approve and send.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: OpenAI business pricing page; OpenAI help centre articles on Projects, data controls and the custom GPT retirement and migration FAQ.