Seven ChatGPT prompts cover a makeup artist's paperwork: an enquiry reply that asks the right questions, a bridal quote, a party and travel quote, a day-of timeline, a multi-person timeline, a kit list per job, and a restock check. Fill in brackets, never real client data, then correct the arithmetic and any invented details.
The prompts below are written to be pasted as they stand. Each comes with a sample output so you can see what a good result looks like and what usually goes wrong. Almost every problem has the same shape: ChatGPT fills a gap you left with something plausible, such as a start time, a price or a product. So the fixes are mostly about closing gaps in the prompt, not editing the output.
Set up once: a standing context so every prompt starts informed
Before the seven prompts, spend ten minutes writing a context block you reuse. In a ChatGPT project (available on the free tier with limits, and with fuller storage on paid plans such as Plus at $20 a month) paste it as the project instructions. In a plain chat, paste it above each prompt.
I am a freelance makeup artist. I work on weddings, parties and photo shoots.
Voice: friendly, direct, professional, British spelling, no exclamation marks.
My rates: Bridal makeup $[x]. Bridesmaid or guest $[y]. Trial $[z].
Hair styling add-on $[a]. Travel: free within [n] miles, then $[b] per mile.
Early start fee (before 7am): $[c]. Deposit: [n]% to secure the date,
non-refundable. Balance due [n] days before.
Never invent prices, products, start times or policies. If something is
missing, ask me or write [CHECK].
The final two lines do most of the work. Telling the model to write [CHECK] instead of guessing turns missing information into something you can see and search for, rather than a confident wrong number sitting in a quote. If you want a rate card to paste into the block, the bridal price list template builds one.
Prompt 1: reply to a first enquiry and ask for what you need
The first message decides whether the bride keeps talking to you. A reply that quotes a price immediately usually loses to one that shows you understand the date, then asks for four details. The prompt:
Write a reply to this enquiry. Thank her, confirm I do [bridal makeup],
say my availability for [date] is [available / tentative / unavailable],
and ask for: number of people needing makeup, ceremony time, venue
location type (home, hotel, outdoor), and whether she wants a trial.
Do not quote a price yet. Under 120 words.
Enquiry: [paste the message with names removed]
A typical sample output (illustrative): "Hi [first name], thank you for getting in touch, and congratulations. I'd be glad to do your makeup on [date] and I currently have that date free. To put together a quote, could you tell me how many people need makeup, the ceremony time, whether you'll be getting ready at home, a hotel or another venue, and whether you'd like a trial beforehand? Once I have those I'll send a full breakdown. Warm wishes, [your name]."
What to fix: availability is the usual error. If you left it as a bracket, the model may write "I'd love to" in a way that sounds like a confirmation. Check the sentence says "currently free" rather than "booked". Also cut any sentence that hints at pricing. The point is to get the facts you need to price correctly, which qualifying wedding leads by date, budget and party size explains in more depth, and how wedding suppliers reply to enquiries in minutes shows the same reply loop for other trades.
Prompt 2: a bridal quote the bride can read in one minute
A quote that is a block of text gets skimmed and queried. A short itemised one gets accepted. The prompt gives the arithmetic to the model but asks for a layout you can check at a glance:
Write a quote as a short email followed by an itemised list.
Use my standing rates. Details:
- Date: [date], getting ready at [venue type], ceremony [time]
- Bride: bridal makeup + [trial yes/no]
- Bridal party: [number] bridesmaids, [number] mothers
- Hair styling for: [number] people
- Travel distance: [n] miles
Show each line, quantity, unit price and line total, then a grand total,
the deposit amount and the balance due date. Add two sentences on what
happens if the date changes. Do not add items I have not listed.
A worked example (illustrative figures): bridal makeup $220, trial $90, three bridesmaids at $95 each, two mothers at $95 each, hair for two people at $60 each, 25 miles of travel with the first 15 free and $1.20 a mile after. The sum: $220 + $90 + $285 + $190 + $120 = $905, plus 10 chargeable miles at $1.20 = $12, giving $917. A 30% deposit is $275.10, and the balance is $641.90.
What goes wrong: language models are unreliable at multi-line arithmetic. Ask for the lines, then check the total with your phone calculator every time. If the sample output reads "Total: $929" and your sum says $917, the quote has a mistake that costs you $12 or embarrasses you in front of the client. A second common fault is an invented extra such as "complimentary lash application"; delete it. The instruction "do not add items I have not listed" reduces this but does not stop it, so read the list against your own before sending.
What the finished quote email looks like
Putting the pieces together, the version a bride receives, once you have corrected the model's draft, is short enough to read on a phone (illustrative):
Hi [first name], thank you for the details. Here is your quote for [date]. Bridal makeup $220. Trial $90. Three bridesmaids at $95 each, $285. Two mothers at $95 each, $190. Hair styling for two at $60 each, $120. Travel, 10 chargeable miles, $12. Total $917. To secure the date I take a 30% deposit of $275.10, and the balance of $641.90 is due 14 days before. If the date needs to change, I will move you to the new date for free if I am available; the deposit is not refundable if I am not. Let me know if you would like to go ahead, and I will send the booking form.
Notice what it leaves out: no adjectives about how magical the morning will be, and no list of services you did not agree. Brides compare several suppliers, and a quote that is easy to scan and complete tends to feel more trustworthy than one that is long. If you are tempted to let the model pad it, ask for it shorter instead.
A mistake that shows up often: the quote that quietly changes the deal
A realistic failure: you ask for a quote for a bride and three bridesmaids, and the model returns a clean table that includes "false lashes for all" because lashes appeared in an example you pasted earlier in the same chat. The bride accepts, and now you have promised lashes to four people at no charge. The remedy is mechanical, not clever. Start a fresh chat for every quote, or use a project so old examples do not leak in, and compare every line of the output against your own list before sending. Ten seconds of checking is cheaper than a $60 give-away.
A follow-up when the quote goes quiet
Many quotes are not declined; they are simply not answered. A follow-up three or four days later, kept to two sentences, recovers a share of them. Ask the model for a version that mentions the date, reminds them the date is not held until the deposit arrives, and offers to answer any questions, with no pressure and no invented scarcity such as "only two dates left" unless it is true. Read it for tone: a reminder that sounds like a debt letter loses the booking. If a week passes with no reply, one last short message that says you will release the date is polite and usually gets an answer either way.
Prompt 3: a party, shoot or event quote with a minimum and travel
Non-bridal bookings have different economics: shorter windows, many people, sometimes several locations. The prompt adds a minimum charge and a time cap, which stop a vague "just a few people, quick touch-up" request turning into a four-hour unpaid morning.
Write a quote for a non-bridal job. Details: [event type], [date],
[number] people, [start time], [location type], hair needed: [yes/no].
Apply my minimum booking of $[n] and my hourly overrun rate of $[n] per
30 minutes. State that the quoted time is [n] minutes per person for a
full look and [n] minutes for a quick refresh. Offer two options:
Option A (full looks) and Option B (quick refresh for some of the group).
Keep it under 150 words plus the list.
The sample output usually offers a tidy two-option comparison. In an illustration: a birthday group of six, four wanting full looks at 45 minutes and two wanting a 20-minute refresh. Option A, all full looks, is six x 45 = 270 minutes, which is four and a half hours. Option B, four full and two refreshes, is 180 + 40 = 220 minutes. Showing the group both durations helps them choose, and the time matters to you more than the price: a 270-minute booking blocks a day.
What to fix: the model tends to round times and quietly drop your overrun rule. Check the quote still says what happens if the group runs late. Also look for tone: some outputs sound like an agency brochure, with phrases like "tailored experience". Replace them with how you would say it.
Prompt 4: a day-of timeline built backwards from the ceremony
The single most useful prompt in this list. A timeline fails when it is built forwards ("start at 6, then 6:45...") because the ceremony time is fixed and everything else must fit behind it. The prompt forces backwards planning:
Build a day-of timeline working BACKWARDS from the ceremony time.
Ceremony: [time]. Photographer arrives to shoot getting-ready at [time].
Venue is [n] minutes' drive from the getting-ready location.
People: bride (needs [n] minutes), [n] bridesmaids ([n] minutes each),
[n] mothers ([n] minutes each). One artist, hair done by [me / someone
else / not needed]. Add a 30-minute buffer before the car leaves and
15 minutes for my own setup. Show times in a table: start, finish,
person, notes. Flag anything that starts before [time] as an early start.
A worked example (illustrative): ceremony 2:00pm, car leaves 12:45pm after a 30-minute buffer, and the bride needs 75 minutes. Work backwards: bride finishes 12:15, starts 11:00. If the bridesmaids need 40 minutes each and there are three, that is 120 minutes, so they start 9:00 and finish 11:00. Two mothers at 40 minutes each is 80 minutes, so they go first, 7:40 to 9:00, with setup at 7:25. The model should produce this chain. If you gave it a different arrangement, the output changes accordingly.
What to check, every time: first, whether the total of all durations plus buffers actually fits before the car leaves; models sometimes finish the last person after the departure time. Second, whether the earliest start is the one you agreed in the quote, because it triggers an early start fee. Third, whether the table has a row for "bride's touch-up before photos", which brides often forget and photographers expect. A wedding planner's AI workflow shows how your timeline slots into theirs.
Adapting the timeline for a trial, a shoot or a long-wear job
The same backwards logic works for other jobs by changing the anchor. For a photo shoot, anchor on the first frame: if the photographer wants the model on set at 10:00 and you need 60 minutes plus 15 of setup, you start at 8:45 at the latest, and a second look halfway through the day needs a gap in the table for a change. For a trial, anchor on the end: a bride who has a hair appointment at 2:00 needs to leave you by 1:30. For a long-wear job such as a 12-hour wedding day, add a touch-up slot at the point where the photos start, since a look that was fine at 11 may need blotting and a lip refresh by 4.
Add one line to the prompt for each: "anchor on the first photograph at [time]", "anchor on the end time of [time]", or "include one touch-up slot at [time] lasting 15 minutes". The model follows these well because they are concrete. It follows vague requests, such as "make sure there is enough time", badly, because it has no number to hold on to.
Prompt 5: one timeline for a team of two artists and a bigger party
Larger parties are where a plain timeline breaks. With two artists, the order changes: both can work in parallel, but the bride must not wait on a stylist who is still finishing a mother. The prompt asks the model to assign people to artists and to point out the bottleneck.
Two artists, [name A] and [name B]. People and minutes: bride 75,
[n] bridesmaids at 40, [n] mothers at 40, [n] flower girls at 15.
Ceremony [time], car leaves [time]. Create a table by artist, with
start and finish times. Put the bride with the more experienced artist
[name A]. Tell me the earliest start and which artist finishes last.
Warn me if anyone finishes less than 20 minutes before the car leaves.
A short illustration: ten people (bride, five bridesmaids, two mothers, two flower girls). Artist A does the bride (75) plus two bridesmaids (80) = 155 minutes. Artist B does three bridesmaids (120), both mothers (80) and two flower girls (30) = 230 minutes. The imbalance is obvious, so the right next step is to move the two flower girls to Artist A, giving 185 and 200 minutes, a gap of 15 minutes instead of 75. That rebalance is something to ask the model for explicitly: "balance the load so both finish within 15 minutes of each other".
What to fix: the model assumes both artists are equally fast. If one is slower on a particular look, such as a dramatic smoky eye, say so in the minutes. And ask it to leave a gap for lash application or a mid-session break, which it will not add on its own.
Prompt 6: a kit list for this job, not a generic one
Most kit lists online are generic, and the real value is a list that changes with the job. The prompt takes the variables that actually change what you carry: the location, the season, the number of skin tones, whether it is outdoors and the number of people.
Create a kit checklist for this job. Group it: skin prep, base,
eyes, cheeks and lips, lashes and brows, tools and hygiene, comfort and
emergency, admin. Job: [n] people, [indoor/outdoor], [season],
[hot/cold/humid], shade range needed: [from light to deep], includes
[trial / photo shoot / long wear for 12 hours]. Mark with * anything
that needs restocking from my standard stock list: [paste list].
Do not name specific brands unless I list them.
A sample result for an outdoor summer wedding with six people (illustrative) will usually add items a basic list omits: blotting papers, a setting spray for long wear, a small handheld fan, tissues and cotton buds for tears, double-sided tape for lashes, a cooler bag for creams that melt in a hot car, and a spare set of disposable applicators for each person. For a cold winter morning venue the same prompt should swap in richer moisturiser, lip balm under lipstick and a heater or hot drinks for yourself.
What to fix: the model will happily add products you do not own. Since the prompt says not to name brands, the list should be generic, but check for items such as "waterproof mascara" if you never use it. Also delete any item that is about skin conditions or treatments outside your scope. Hygiene lines are the safest place to be conservative: sanitise tools between clients, use disposable applicators for lips and mascara, and keep a patch-test or allergy note in your admin folder, not in a chat tool.
Example: the same kit list for a daytime wedding and an evening shoot
Run prompt 6 twice and compare, because the contrast shows how much the variables matter. For a summer afternoon wedding with eight people at an outdoor venue, expect the output to stress sweat control and a longer-lasting base, spare lashes and glue, a handheld fan and a bag of blotting papers, and enough sets of disposable brushes for eight faces. For an evening studio shoot with two models, the list should shrink the quantities and grow the specialist items: highlighter and contour for the lighting, a lint roller for clothing, and a second look kit so you can change a face quickly between setups.
If the two lists come out nearly identical, your prompt was too thin. The fix is to add the one fact the model needs, such as "outdoors, 28 degrees, four hours between makeup and photographs". That is a detail a generic article would never have, and it is the reason to use a prompt rather than a downloadable checklist.
Setting up an emergency box from the output
Prompt 6 asks for a comfort and emergency group, and it is worth turning it into a permanent small box that travels to every job. A reasonable result includes plasters, safety pins, double-sided tape, tissues, a sewing needle and thread, a lint roller, hand sanitiser and a small mirror. Several of those have nothing to do with makeup but have saved many wedding mornings. Print the group as a one-page laminated card and keep it in the kit lid, so you check it by sight without opening your phone, which also keeps client messages out of your sight at the wrong moment.
Prompt 7: a restock check after the job
The last prompt turns the job into the next job's prep. After each booking, list what you used and let the model compare it with your standing stock minimums:
I just finished a job. Items used (with rough quantities): [list].
Here is my stock list with minimum levels: [paste]. Tell me which items
are below minimum, what to order first, and one thing I should check
before the next booking on [date]. Format: a short list I can paste
into my phone.
A short illustration: your minimum for disposable lip applicators is 40 and you started with 55 and used 18 on a ten-person job. The output says 37 left, which is below minimum, so order. If you also used the last of a particular foundation shade, it flags that first because it is the item you cannot buy locally on a Saturday morning. The check that matters is the date of the next booking, because some restocks have delivery times.
When a date changes: a short reschedule prompt
Not one of the seven prompts, but a useful companion: a change of date is where quotes go stale, because your availability, early-start fee and travel may all be different. Paste the original quote, the new date and your calendar status, and ask for a one-paragraph reply that confirms what changes and what does not, and a rebuilt total if anything differs. Then re-run the timeline prompt with the new ceremony time rather than adjusting the old table by hand, since moving one time in a table usually breaks a buffer elsewhere.
Keep client details out of the prompts
All seven prompts use placeholders, and that is deliberate. Names, phone numbers, venue addresses, skin conditions, allergies and medical notes do not belong in a general chat tool. Write the prompt with [bride] and [venue], paste the output into your own document, and fill in the details there. If you want ChatGPT to learn from your past quotes, remove the names first and keep only the structure. The tool will not mind the difference; your clients may.
If you mention in your terms how you handle information, how to mention AI use in client contracts and proposals gives wording you can adapt, and it is fine for a one-person business to be plain about it.
Which of the seven to start with
If you only adopt two, take prompts 2 and 4. The quote is where money is won or lost, and the timeline is where the day itself runs smoothly or does not. Prompt 1 comes next because it cuts your reply time, then 6 and 7 once you are doing three or more jobs a month. Prompts 3 and 5 matter only if you take on groups or work with a second artist. Treat the sequence as a trial: use one prompt for two weeks, note what you corrected, and adjust the standing context before adding the next.
How to check a month of use is working
- Time the first reply: from reading an enquiry to sending a good answer. Aim for under five minutes, from perhaps fifteen.
- Count quote errors you caught before sending. If it is more than one in five, your standing context is missing something; add it.
- Track conversion: how many quotes turned into deposits. If it has not changed, the wording may be fine, but the price or availability may be what decides it.
- Compare planned and actual finish times on three jobs. If you are regularly twenty minutes late, your minutes-per-person figures are too optimistic; change them in the context block, not in each prompt.
None of this replaces your judgement about a bride's nerves, a venue's lighting or a hot day. It replaces the typing and the arithmetic, which is where most of the evening's admin goes.
Questions makeup artists ask about using ChatGPT
Should I paste a client's name and details into ChatGPT?
Leave them out. Use [bride], [venue] and [date] placeholders, then add the real details in your own document afterwards. The prompts below work without any personal information, and skin conditions or allergies are the last thing to put into a chat tool.
Can ChatGPT set my prices?
It can calculate from rates you give it, but it cannot know your market, your costs or your minimum. Give it your price list and let it do the arithmetic and wording. Always re-add the totals yourself before sending a quote.
Which ChatGPT plan do I need?
The free tier handles all seven prompts. A paid plan such as Plus at $20 a month mainly raises limits and adds projects, where you can store your price list and policies once instead of pasting them every time.
Will clients know a quote was written with AI?
They will notice generic wording, not the tool. If every quote is in your voice, accurate and specific to their date and look list, the source is irrelevant. Read each one aloud and cut any sentence you would not say to the bride's face.
Further reads
- HoneyBook AI vs ChatGPT for Wedding Planners: Proposals and Emails — When a studio tool's AI beats a plain chat prompt.
- Wedding Photographer Timeline Template: A Day Schedule Built With AI — The photographer's side of the same wedding-day schedule.
- AI Mistakes That Put Couples Off Your Wedding Business — AI slips that put couples off a wedding supplier.
- ChatGPT Prompts for Pet Sitters and Dog Walkers That Win Clients — Another solo-trade prompt set with different sample outputs.
- Florist Wedding Quote Template: Price Stems, Labour and Hire With AI — Another wedding quote sheet with labour and travel lines.
- How to Write Wedding Flower Proposals With AI — Proposal structure another wedding supplier uses.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: OpenAI ChatGPT pricing page (checked October 2026). The prompts and sample outputs are illustrative and contain no real client data.