Create an AI Dress Campaign Kit

Create a truthful dress launch campaign kit with ChatGPT and Canva.

Introduction

30 Second Summary

A beautiful dress can disappear in a crowded feed when the message feels generic. A strong launch gives shoppers a clear reason to pause.

In this project, you will turn verified facts from the Maison Plissé website into a truthful dress-launch campaign using ChatGPT and Canva. You will prepare the finished campaign kit for adult review before anything is published.

What You'll Build

When you finish, you can show an adult a polished Maison Plissé product graphic paired with a truthful caption and a clear campaign goal.

By the end of this project, you'll have:

A Product Truth Sheet that separates current website facts from ideas and unconfirmed claims.
A focused campaign brief with a target-customer hypothesis, supported benefits, a three-word brand voice, campaign hooks, and a clear call to action.
A polished PNG campaign graphic with a matching caption and an adult handoff note for review before publishing.
Secret Mission: A second campaign version that changes only the hook so an authorized adult can compare two messages fairly.

Are there any prerequisites?

You need existing ChatGPT and Canva access on Windows plus the public Maison Plissé website.

You also need confirmed permission from a parent or legal guardian before you begin.

Before We Start

A clear commitment keeps your campaign grounded in one current Maison Plissé product. It also keeps approval and publishing decisions with an authorized adult.

Set Up a Safe Campaign Workspace

A safe workspace keeps private customer data out of your campaign draft. A live public source keeps product claims tied to information that customers can verify.

In this step, you will prepare a protected ChatGPT conversation for your campaign. You will also open a Canva design that follows the authorized adult's decision about photo reuse.

In this step, get ready to:

Establish a safe ChatGPT campaign conversation.
Verify one current Maison Plissé product with an authorized adult.
Open a Canva social-post design for the approved visual path.

Open the campaign tools

The Maison Plissé website is the public source for this campaign. On Windows, a current version of Microsoft Edge or Chrome can load the website plus your existing campaign tools.

Press the Windows key to open Windows search.
Type Microsoft Edge or Chrome into the search field.
Press Enter to open the selected browser.

You should see a browser window ready for your campaign tabs.

Enter https://maisonplisse.nl in the browser address bar.

You should see the public Maison Plissé website. This live website is where you will confirm that your chosen product is still current.

Open a new browser tab.
Load ChatGPT through the existing access you confirmed before this project.

You should see either your ChatGPT home screen or a sign-in screen.

Sign in with your existing ChatGPT account if prompted.
Start a blank ChatGPT conversation.
Set the adult-review boundary by sending this as the conversation's first message:

AI draft. Adult review required before publishing.

Why Start With This Message?

This message labels every response as an AI draft. It keeps review plus publishing decisions with the authorized adult.

Keep customer names or addresses out of the conversation.
Keep orders or payment information out of the conversation.
Keep private messages out of the conversation.
Keep credentials out of the conversation.
Keep unpublished business information out of the conversation.

Confirm that the adult-review boundary is the first user message in the conversation.

Before you send the workspace check, predict what response you expect from this new campaign conversation.

Test the campaign conversation by sending:

Campaign workspace ready

What Does This Check Prove?

A response confirms that the campaign conversation can receive your messages. It creates a visible checkpoint before any approved product information enters the draft.

You should see a ChatGPT response beneath your message. That confirms your protected drafting space is working.

Conversation Not Responding?

Confirm that ChatGPT shows you as signed in.
Reload the ChatGPT browser tab.
Send the workspace check again after the conversation reloads.

Help me troubleshoot why my existing ChatGPT conversation is not responding.

Confirm the approved product path

Your campaign commitment already names a Maison Plissé product. The live listing now confirms whether that product is still available for this campaign.

Your commitment may name Noir – Black Lace Jumpsuit or Aurelia – Champagne Sequin Evening Dress. The live website decides whether either product is still current.

Return to the Maison Plissé website tab from earlier.
Find the product named in your campaign commitment on the live listing.

You should see the selected product on the current website. Its exact name, availability, plus price remain subject to verification before they become campaign claims.

Select another product from the live listing if your original product no longer appears.
Update your campaign commitment if your selected product changed.

The product in your commitment should now match one product shown on the live Maison Plissé website.

Why Recheck the Live Listing?

Product names can change. Availability can change.

Prices can change too. A live check prevents old information from becoming a new campaign claim.

The authorized adult now decides which public facts are approved for drafting. This review keeps business decisions with the person responsible for the campaign.

Show the selected live product page to the authorized adult.
Ask the authorized adult to approve the public product facts you may use.

You should receive a clear decision about which public facts are approved for your campaign drafts.

Photo permission can be tricky because access to an image does not automatically grant reuse rights. The authorized adult must make this decision before the image enters Canva.

Ask the authorized adult to approve one specified business-owned product photo.
Select the text-only Canva fallback if photo reuse is not approved.

Your campaign now has one approved visual path. You can use the specified business-owned photo or continue with a text-only design.

Open the Canva campaign design

A social-post design gives your campaign a visual workspace. The adult's photo decision determines whether this workspace uses the specified product photo or the text-only fallback.

Open a new browser tab.
Load Canva through your existing access.

You should see either the Canva home page or a sign-in screen.

You should see your Canva home page without creating another account.

Start a new social-post design from the Canva home page.

You should see the Canva editor with an open social-post canvas. The canvas is ready for the adult-approved photo path or the text-only fallback.

Canva Editor Not Loading?

Switch to a current version of Microsoft Edge or Chrome if Canva reports that your browser is unsupported.
Reload Canva in the current browser.
Confirm that you are using your existing Canva account.

Help me troubleshoot why the Canva social-post editor is not loading.

Before you check both tools, predict which two visible signs will prove that your campaign workspace is ready.

Switch back to the ChatGPT browser tab.
Confirm that the conversation begins with AI draft. Adult review required before publishing..
Confirm that a response appears beneath Campaign workspace ready.
Switch back to the Canva browser tab.
Confirm that the social-post canvas is open inside the editor.

You should see the protected ChatGPT conversation plus the open Canva editor. Your Canva canvas should match the approved photo path or remain ready for the text-only fallback.

One Workspace Still Missing?

Return to the correct browser tab if one tool is hidden.
Reload the tool that did not finish loading.
Confirm that the ChatGPT conversation still starts with the adult-review boundary.

Help me identify which part of my safe campaign workspace is incomplete.

Your campaign workspace now protects private information while keeping the authorized adult in control. Next, you will test what happens when ChatGPT receives a vague dress prompt.

Create a Product Truth Sheet

Your campaign workspace is ready. ChatGPT still lacks the verified context for your selected Maison Plissé product.

This step exposes that gap with a deliberately vague prompt. You will turn the weak result into a Product Truth Sheet grounded in the live listing and adult-approved information.

In this step, get ready to:

Generate a deliberately generic caption draft.
Identify at least three missing or invented product details.
Complete a Product Truth Sheet using verified information.

Test the vague prompt

A prompt supplies the context that guides an AI response. This test uses only a broad request so you can judge what the response knows from those words alone.

Before you send the prompt, predict which product details ChatGPT can know from these words alone.
Switch back to the ChatGPT campaign conversation from earlier.
Send the deliberately vague prompt by copying this text:

Write three social-media captions for a dress.

What does this prompt test?

The request gives ChatGPT a format and a broad product category. It provides no verified details about your selected Maison Plissé product.

Any product-specific detail in the response lacks support until you match it to the live listing or adult-approved information.

Wait for ChatGPT to return three captions.
Keep the response in the conversation as your weak first draft.

You should see three broad captions with little connection to the selected product. Any specific claim may also lack evidence.

No response from ChatGPT?

Check that the ChatGPT conversation from earlier is still active.
Resend the exact prompt shown above.

Ask for help if the conversation does not return the three captions.

Audit the weak draft

The live product listing is your evidence source. A caption detail is usable only when the listing or adult-approved information supports it.

Return to the selected Maison Plissé product listing from earlier.
Compare each caption with the exact product name shown on the listing.
Verify every visual description against the listing images or text.
Verify every material or construction claim against the listing.
Check whether sizes or options are missing or unsupported.
Check whether the price or availability is missing or unsupported.
Check whether any shipping or return wording is missing or unsupported.
Mark every unsupported detail as unusable.

Your review should identify at least three details that are missing or invented. The intended shortfall proves that the vague prompt cannot produce a trustworthy product campaign.

Build the Product Truth Sheet

The Product Truth Sheet creates a boundary around the claims your campaign may use. Anything outside that boundary stays out of the final campaign.

Create a blank local document using your usual notes app.
Build the Product Truth Sheet by copying this template into the document:

# Product Truth Sheet

Exact product name:
Source URL:
Date checked:
Verified visual details:
Verified material or construction details:
Verified sizes or options:
Verified current price or omit it:
Verified availability:
Verified shipping or return wording:
Facts that must not be invented:

How does this truth sheet work?

Each verified entry comes from the live listing or adult-approved information.
The value Not confirmed labels information that is absent or unapproved.
The final field records claims that future AI drafts must avoid.

Enter the exact current product name shown on the live listing.
Paste the selected product page URL into the source field.
Record the date when you checked the listing.
Fill each verified field using the public listing or adult-approved information.
Record the current price only when the live listing shows it.
Mark the price as omitted when you choose to exclude it from the campaign.
Enter Not confirmed for every absent or unapproved field.
Save the completed Product Truth Sheet document.

Before your final review, predict whether every product claim can be traced to the live listing or adult approval.
Audit each populated field against its supporting source.
Remove every detail that lacks support.
Confirm that every absent or unapproved field says Not confirmed.

You should now have a complete truth sheet with no unsupported product claims. That is the key safeguard in place for every campaign draft that follows.

Unsure whether a detail is verified?

Use Not confirmed when the listing does not clearly support the detail.
Ask the authorized adult to confirm any business information that does not appear publicly.

Get help separating verified facts from assumptions.

Your Product Truth Sheet now gives the campaign a reliable evidence base. Next, you will use it to build a fact-checked campaign brief.

Build a Truth-First Campaign Brief

Your Product Truth Sheet gives ChatGPT the evidence that the weak captions lacked. The draft still needs a clear strategy.

In this step, you will turn those verified facts into a focused Campaign Brief. A claim audit will keep every product statement tied to evidence.

In this step, get ready to:

Generate campaign options from the Product Truth Sheet.
Select the components for a fact-safe Campaign Brief.
Audit every factual product claim against its source.

Generate evidence-based campaign options

A customer hypothesis describes who may care about the selected product. It stays labeled as an assumption until customer evidence supports it.

Switch back to the ChatGPT campaign conversation from earlier.
Paste your completed Product Truth Sheet into the message box.
Add the truth-first campaign instructions directly below the Product Truth Sheet by pasting this prompt:

Act as a beginner-friendly ecommerce marketing assistant.
Use only product facts from my Product Truth Sheet.
Treat Maison Plissé website copy only as a brand-tone reference, not as permission to invent product claims.
Label customer assumptions as hypotheses, not facts.
Create:
1. One short target-customer hypothesis
2. Three customer benefits that are directly supported by product facts
3. A three-word brand voice
4. Three campaign hooks
5. One simple call to action
Do not invent discounts, materials, sizes, stock, reviews, delivery dates, popularity, sustainability claims, or guarantees.

How Does This Prompt Protect the Brief?

The instruction to use only the Product Truth Sheet limits factual product statements to approved evidence.
The hypothesis label keeps customer assumptions separate from product facts.
The final restriction blocks common marketing claims that the source may not support.

Submit the combined Product Truth Sheet and campaign prompt.

You should see one target-customer hypothesis plus three supported benefits. You should also see a three-word voice plus three hooks plus one call to action.

That first strategic draft is ready, with the Product Truth Sheet controlling what ChatGPT can claim.

Output Includes Unsupported Details?

Compare each new product detail with the Product Truth Sheet.
Remove any detail that has no matching line.

Ask for help to identify unsupported claims in my campaign options.

Select the Campaign Brief

A three-word brand voice guides how the campaign sounds. Each chosen campaign element must still point back to the Product Truth Sheet.

Choose one target-customer hypothesis from the generated options.
Keep the label HYPOTHESIS beside the chosen customer statement.

Your audience statement now communicates a possible customer without presenting that assumption as a fact.

Select one product benefit that is directly supported by the Product Truth Sheet.

The selected benefit should describe customer value that the verified product details can support.

Keep all three fact-safe campaign hooks in the Campaign Brief.
Mark one hook as the selected approved hook.

The brief now preserves three usable directions while identifying the message that moves forward.

Confirm that the selected brand voice contains exactly three words.
Keep one simple call to action that does not promise an unverified outcome.

What Makes the Brief Fact-Safe?

The customer statement remains a hypothesis. The product benefit comes from verified facts.

The hooks can be creative without adding product claims. The call to action gives the customer a next step without inventing urgency.

Rewrite any wording that does not fit Maison Plissé.
Keep the chosen pieces together under the heading Campaign Brief in the same conversation.

Your Campaign Brief should now contain one labeled hypothesis plus one supported benefit. It should also contain a three-word voice plus three fact-safe hooks plus one selected hook plus one simple call to action.

Audit every campaign claim

A claim audit tests each factual statement against its source. It also keeps customer assumptions visibly separate from verified product information.

Paste this audit prompt directly below the Campaign Brief in the same conversation:

List every factual product claim in this Campaign Brief. For each claim, quote the matching line from the Product Truth Sheet. Mark unsupported claims as REMOVE and customer assumptions as HYPOTHESIS.

What Does the Audit Check?

Each factual product claim needs a matching quotation from the Product Truth Sheet.
Each unsupported statement receives the label REMOVE.
Each customer assumption receives the label HYPOTHESIS.

Before you submit the audit, will every factual product claim have an exact match in the Product Truth Sheet?

Send the claim-audit prompt.

You should see every factual claim paired with a quoted Product Truth Sheet line. Any unsupported item should appear with the label REMOVE.

Delete every item marked REMOVE from the Campaign Brief.
Rewrite an unsupported item only when verified wording from the Product Truth Sheet can replace it.

The revised Campaign Brief should now contain no unsupported product claims.

Keep the label HYPOTHESIS on every customer assumption.

Before you rerun the audit, will any factual claim remain without a matching Product Truth Sheet line?

Run the final check by sending the same claim-audit prompt again.

You should see every factual product claim mapped to a quoted Product Truth Sheet line. You should see customer assumptions labeled HYPOTHESIS with no remaining REMOVE items.

You have now closed the evidence gap that weakened the first captions. The Campaign Brief is ready to guide the visual campaign.

Audit Still Finds Unsupported Claims?

Check that the complete Campaign Brief appears directly above the audit prompt.
Confirm that every claimed product detail appears in the Product Truth Sheet.
Replace any unsupported wording with a verified statement.

Ask for help to resolve the remaining unsupported campaign claims.

Your truth-first Campaign Brief is locked in. Next, you will turn its approved hook plus supporting message into a Canva graphic plus matching caption.

Design, Review, and Export the Campaign

Your fact-checked Campaign Brief links a customer hypothesis to one evidence-backed benefit. The selected hook comes from your Product Truth Sheet.

A strategy becomes useful when it turns into a clear visual with matching copy. Human review keeps the final publishing decision with an authorized adult.

You will use ChatGPT to shape the approved brief into final copy. Canva will turn that copy into a review-ready graphic.

In this step, get ready to:

Generate a final campaign package from the approved Campaign Brief.
Build a Version 1 Canva graphic with fact-safe copy.
Export the PNG and prepare an adult handoff note.

Turn the approved brief into final copy

The claim audit has already removed unsupported details. The final prompt keeps the headline and caption inside that approved boundary.

Return to the ChatGPT campaign conversation from earlier.
Find the approved Campaign Brief below the completed claim audit.
Generate the final campaign package by sending this prompt:

Using only this approved Campaign Brief, create:
1. One short headline for the graphic
2. One concise social caption
3. One call to action
4. A final checklist of factual claims that need adult verification
Do not add new product facts, discounts, urgency, reviews, shipping promises, or guarantees.

How does this prompt protect the campaign?

The opening instruction limits the draft to material that has already passed your claim audit.
The four requested parts keep the graphic copy aligned with the social caption.
The final instruction blocks common marketing claims that lack approved evidence.

Confirm the response contains one short graphic headline.
Confirm the response contains one concise social caption.
Confirm the response contains one call to action.
Confirm the response contains a factual-claim verification checklist.
Compare every factual product claim with the approved Campaign Brief.
Ask ChatGPT to remove any content that introduces a new or unsupported claim.
Keep the cleaned final package in the campaign conversation.

You should now see four clearly labeled parts. Every factual statement should trace back to the approved Campaign Brief.

Seeing new claims?

Point ChatGPT back to the approved Campaign Brief. Remove any line that adds a discount, urgency, review, shipping promise, guarantee, or unverified product fact.

Ask for help with the comparison: help me identify any claim in this final campaign package that is missing from my approved Campaign Brief.

Build and export Version 1

The Canva design from earlier is ready for its approved content. Your adult approval record determines whether the visual uses the business-owned product photo or the text-only fallback.

Return to the Canva social-post design from earlier.

Use the approved product photo

Use only the specific business-owned photo covered by the adult approval record.

Select Upload your image in Canva.
Choose the adult-approved product photo from your Windows computer.
Place the uploaded photo in the social-post design.

Use the text-only fallback

Continue without uploading a product photo. The approved campaign copy becomes the visual focus.

Leave all unapproved product photos out of the design.
Choose a plain background or a free Canva asset.
Reserve most of the canvas for the approved text.

Set the selected approved hook as the largest text in the design.
Add Maison Plissé as the brand name.
Add the approved call to action.
Use only assets that are free in your Canva access.
Set text colors that contrast strongly with the background.
Leave enough empty space around the product or fallback message.
Review the completed design in the Canva editor.

You should be able to read the headline without zooming in. The product or text-only fallback should remain visually clear.

Your Version 1 visual now carries the same verified message as the Campaign Brief.

Design hard to read?

Increase the headline size if it disappears into the layout. Change the text color if the background reduces contrast.

Get help refining the layout: help me improve the readability of my Canva campaign graphic without changing its approved copy.

Use Canva's download controls to select PNG as the file format.
Download the Version 1 design.

The exported graphic should now appear in your Windows Downloads folder.

In the Windows Downloads folder, rename the exported file to maison-plisse-launch-v1.png.

You should now see maison-plisse-launch-v1.png in the Downloads folder.

Cannot find the PNG?

Check your browser's recent downloads for the Canva export. Confirm that you downloaded the design in PNG format before renaming it.

Get help locating the export: help me find and rename my Canva PNG in the Windows Downloads folder.

Prepare the adult handoff

A handoff note gives the adult reviewer the context behind the finished graphic. It also names the result they can measure after publishing.

Return to the ChatGPT campaign conversation from earlier.
Start a new message with the heading Adult handoff note.
Add the final social caption.
Add the PNG filename maison-plisse-launch-v1.png.
Add the target-customer HYPOTHESIS from the Campaign Brief.
Add the campaign goal.
Add the Product Truth Sheet source URL.
Add the date when you checked the source.

Why choose one metric?

One metric gives the authorized adult a clear result to monitor after publishing. Product-page visits, inquiries, or sales can connect the campaign to a specific business outcome.

Add one suggested post-publication metric.
Add an instruction to recheck image rights.
Add an instruction to recheck every product fact.
Add an instruction to recheck the price if the graphic or caption shows one.
Add an instruction to recheck current availability.
Add an instruction to recheck the product link.
Add an instruction to recheck the publishing account.
Send the completed handoff note into the campaign conversation.

Before you open the export, do you expect every required element to remain readable outside Canva?

Open maison-plisse-launch-v1.png from the Windows Downloads folder.

You should see the adult-approved product photo or the text-only fallback. The selected approved hook should be readable.

The name Maison Plissé should be readable. The approved call to action should also be readable.

Show the opened PNG to the authorized adult reviewer.
Show the adult reviewer the final caption and handoff note.
Ask the adult to approve the graphic and caption before publishing.
Wait for explicit adult publishing approval before anything goes live.

Publishing affects a real business, so pausing here protects the brand. The authorized adult keeps control of the final publishing decision.

That is the full campaign handoff ready. Final publishing control stays with the authorized adult.

Secret mission

Create a Controlled Hook Test

One campaign shows a finished idea. A second controlled version helps an authorized adult compare two approved hooks while every other campaign element stays consistent.

Clean Up Your Resources

Your campaign uses existing ChatGPT and Canva access, so there are no additional paid resources or ongoing costs.

Decide whether to keep the campaign workspace for another Maison Plissé product, pause before publishing, or remove unwanted drafts while retaining adult-approved files.

Resources you used:

One ChatGPT campaign conversation containing the weak draft, Product Truth Sheet, Campaign Brief, claim audit, final package, and adult-review reminder.
Two Canva campaign designs containing the approved Version 1 graphic and controlled Version 2 hook test.
Two PNG exports named maison-plisse-launch-v1.png and maison-plisse-launch-v2.png in your Windows Downloads folder.
One adult handoff note containing the caption, campaign context, source details, verification checklist, and suggested metric.
One controlled hook-test note naming both hooks and the shared comparison metric.

Keep everything running

No action needed. Choose this if you plan to reuse the truth-first campaign workflow for another Maison Plissé product.

Keep the ChatGPT conversation as a record of the Product Truth Sheet, claim audit, and approved campaign copy.
Keep both Canva designs so you can update verified product details or prepare another controlled hook test.
Keep both PNG exports with the adult handoff note so the reviewer can match each graphic to its campaign context.
Keep publishing on hold until an authorized adult rechecks image rights, product facts, any displayed price, availability, the destination link, and the publishing account.

Pause - I'll come back to this later

There are no servers or paid resources to pause. Your saved conversation, designs, PNGs, and notes remain available while publishing stays on hold.

Keep maison-plisse-launch-v1.png and maison-plisse-launch-v2.png in your Windows Downloads folder.
Leave the ChatGPT campaign conversation saved for future fact checks or copy revisions.
Leave both Canva campaign designs saved for future visual revisions.
Store the adult handoff note with the two approved PNGs.
Stop before publishing until the authorized adult is ready to complete the final review.

Delete - I don't want to use this again

Deleting online drafts can feel final. Adult approval defines which campaign files remain after cleanup.

Clean the online drafts:

Review the existing ChatGPT campaign conversation for accidental private information.
Delete the entire ChatGPT campaign conversation with its conversation controls if it contains private information or you do not want to keep the draft.
Confirm the campaign conversation no longer appears in your ChatGPT conversation list.
Review both Canva campaign designs for accidental private information or unwanted drafts.
Delete each unwanted Canva campaign design with the deletion control attached to that design.
Confirm each deleted campaign design no longer appears in your Canva design list.

Clean the downloaded files and notes:

Ask the authorized adult which campaign files are approved to retain.
Return to your Windows Downloads folder in File Explorer.
Delete maison-plisse-launch-v1.png only if the authorized adult says it should not be retained.
Delete maison-plisse-launch-v2.png only if the authorized adult says it should not be retained.
Confirm the PNG filenames remaining in your Windows Downloads folder match the adult's approval decision.
Retain the adult handoff note with every approved PNG.
Delete the adult handoff note only if the authorized adult confirms the entire campaign will not be used.
Retain the controlled hook-test note if the adult plans to compare Version 1 with Version 2.
Delete the controlled hook-test note if the controlled comparison will not proceed.

Your cleanup is complete when no accidental private information remains and every retained file has adult approval.

Nice Work!

You did it! You built a truthful Maison Plissé dress-launch campaign kit using ChatGPT plus Canva with two controlled graphics plus an adult review handoff.

What you learned:

Built a Product Truth Sheet from the gaps exposed by a vague AI prompt.
Created a fact-checked Campaign Brief that turns verified product evidence into customer-focused copy.
Delivered a readable Canva campaign package that an authorized adult can review before publishing or measuring results.
Completed the Secret Mission by creating a controlled hook test where only the headline changes between two campaign versions.

Ready to quiz yourself?