If your product catalog is stored in a PDF spec sheet, website catalog, scrambled Excel file, or supplier document, you do not need to manually retype every product into our CSV format.
You can use generative AI assistants — such as Google Gemini, ChatGPT, Anthropic Claude, or xAI Grok — to extract, calculate, and format your product catalog into our exact bulk CSV structure in under 5 minutes.
Step 1: Choose your AI assistant
All major AI assistants (Google Gemini, ChatGPT, Anthropic Claude, xAI Grok) support data extraction and formatting. Select whichever one you have access to, and use the most capable model available on your plan — newer models generally follow the formatting rules more reliably.
Step 2: Copy the master AI prompt
Gather your source materials (website link, product PDF, or raw Excel file) and the blank upload CSV template from the portal (e.g., WISEcode-Bulk-Upload-Example.csv). Then copy and paste the master prompt below into your AI chat window:
Role: Expert Data Analyst & E-commerce Catalog Specialist.
Task: Extract product data from my source material and format it into the exact CSV structure provided in the attached CSV template file.
Source Material: [PASTE PRODUCT WEBSITES, RAW TEXT, OR ATTACH PDF/EXCEL]
CSV Template Headers: [ATTACH BLANK SAMPLE CSV FILE OR COPY HEADER ROW HERE]
Strict Instructions & Formatting Rules:
1. HEADER MATCH: The output CSV header must EXACTLY match the template headers word-for-word, including asterisks (*) and units in parentheses. Do not rename, reorder, add, or delete columns.
2. REQUIRED FIELDS & ZEROS: Fill in all mandatory fields marked with an asterisk (*). Enter 0 when a value is truly zero — an empty cell means 'missing' and will hold the product in Bulk Review.
3. UPC VALIDATION (CRITICAL):
- Every UPC must be a valid GTIN of 12 to 14 digits with no spaces or dashes.
- You MUST compute or verify the GS1 Modulo-10 check digit (the final digit of each GTIN) so every UPC passes automated validation.
- Preserve leading zeros (treat UPCs as text strings, not numbers).
- Each UPC must be unique within the file — a duplicate UPC on more than one row rejects the ENTIRE file.
4. BRAND MATCH: The Brand value must exactly match one of the registered brand names for my organization, spelled identically. An unrecognized brand blocks the entire file. My registered brands are: [LIST YOUR REGISTERED BRAND NAMES]
5. INGREDIENT LIST & COMMA ESCAPING: Copy the ingredient statement exactly as printed on the label, in descending order by weight, including sub-ingredient parentheses. Enclose ingredient lists, descriptions, and allergen statements in double quotes ("...") so commas do not offset CSV columns.
6. NUMERIC FIELDS: Serving size, prices, calories, and nutrient values must be numbers only (e.g., "14" for serving size, not "14g"). Serving Size Unit must be one of: g, mg, oz, floz, ml, l. Enter Vitamin D in mcg, not IU.
7. STANDALONE VS MULTI-PACK: Treat each individual SKU size (e.g., 3 oz vs 8 oz) as a separate row with its own unique UPC, and leave 'Base Unit UPC' and 'Package Label' blank. For a true multi-unit bundle/case: fill UPC, Product Name, Brand, Base Unit UPC (the single unit's UPC), Package Label (e.g., "6-pack"), and Servings Per Container counted for the WHOLE pack (six bags of 8 servings = 48); leave ingredient, serving size, and nutrition columns blank — the system copies them from the base unit.
8. NFP-EXEMPT PRODUCTS: If a product has no FDA Nutrition Facts Panel, set 'NFP Exception?' = Y and leave the serving and nutrition columns blank. The ingredient list is still required.
9. IMAGE URL: Include a public https front-of-pack image URL when available (optional, but the product cannot reach Complete status without one).
10. FILE LIMIT: Maximum 500 products per file. If the catalog is larger, split the output into multiple CSV files of up to 500 rows each.
Output Requirement:
Output the final complete product catalog as a single plain-text CSV code block that I can copy and save directly as a .csv file.Step 3: Verify UPC check digits and uniqueness
The #1 reason bulk uploads fail is an invalid UPC check digit (a mathematical mismatch on the last digit of the barcode). Ask your AI assistant to run an explicit check before saving:
Please double-check all UPCs in the table. Every UPC must be a valid 12- to 14-digit GTIN: run the official GS1 Modulo-10 check-digit formula on each one and confirm the final digit is mathematically valid. Also confirm that no UPC appears on more than one row. Re-print the updated CSV block if anything was adjusted.
Step 4: Save as CSV (UTF-8) and upload
On Mac (TextEdit)
- Copy the code block generated by the AI.
- Open TextEdit and paste the text.
- Click Format > Make Plain Text (Shift + Command + T). Critical step: this prevents saving as Rich Text.
- Save the file as
catalog_upload.csvand select Use .csv when prompted.
On Windows (Notepad)
- Copy the code block generated by the AI.
- Open Notepad and paste the text.
- Click File > Save As.
- Change Save as type to
All Files (*.*)and set Encoding toUTF-8. - Name the file
catalog_upload.csvand save.
In Excel / Google Sheets
- Paste the content into a blank worksheet.
- In Excel, use File > Save As > CSV UTF-8 (Comma delimited) (.csv). In Google Sheets, use File > Download > Comma-separated values (.csv).
Remember: each file can contain up to 500 products. After upload, complete products import automatically; anything with a gap waits in Bulk Review, where you can correct and re-upload by UPC without starting over.
Troubleshooting common upload errors
| Error / symptom | Cause | How to fix via AI |
|---|---|---|
| "UPC check digit failed" | Excel stripped leading zeros or the last digit does not match GS1 math rules. | Prompt the AI: "Format all UPC values as strict strings of 12–14 padded digits, ensuring leading zeros are preserved and every check digit passes GS1 Modulo-10." |
| Entire file rejected | One of the three file-level blockers: the same UPC appears on more than one row, a brand name does not match a registered brand, or the header row was renamed, reordered, or removed. | Prompt the AI: "Check for duplicate UPCs across rows, confirm every Brand value exactly matches my registered brand list, and re-print the header row exactly as in the template." |
| "Columns shifted or misaligned" | Ingredients or descriptions containing commas were not enclosed in double quotes. | Prompt the AI: "Wrap all text fields containing commas in double quotes and re-output the CSV." |
| "Invalid unit in Serving Size" | Text letters were placed inside a numeric field (e.g., 14 grams). | Prompt the AI: "Separate numeric values into 'Serving Size *' (number only) and 'Serving Size Unit *' (g/mg/oz/floz/ml/l)." |
| Product stuck in Bulk Review | Required cells were left empty. An empty cell means "missing," not zero. | Prompt the AI: "Enter 0 in any required nutrition field whose true value is zero, and flag any fields that are genuinely unknown so I can look them up." |
Need additional support? Attach your AI-generated CSV file and error log to a support ticket, and our onboarding team will assist you directly.
Comments
0 comments
Please sign in to leave a comment.