Ask a chatbot what a wedding costs and it hands back a national average, which is the one number guaranteed not to be yours. Ask it to review a venue quote and it will cheerfully do the math on figures it just made up. The tool is capable; the prompt is the problem.
So we built a set of fourteen prompts that close those two gaps. They are free, they need no signup, and they work in ChatGPT, Claude, or Grok, whichever you already use.
Your numbers go in before the model says a word
Each prompt page has a short form. City, guest count, total budget, month, and on the pages that need them, the specifics: a venue’s food and beverage minimum, a contract’s cancellation terms, a dress budget with alterations inside it. Type your numbers once and every prompt on the site rewrites itself around them, in your browser, without sending us anything.
The part we care most about: they refuse to guess
Leave a required field blank and the prompt does not let the model improvise. Its first reply has to be a numbered list of questions asking for what is missing, and nothing else. A confident budget built on invented figures is worse than no budget, because it looks finished.
The same honesty runs through the answers. The venue quote prompt shows its arithmetic as a ledger you can check line by line against the proposal. The venue shortlist must say whether its list came from live search or from memory, and “unknown” is an accepted answer where an invented capacity is not. The contract review is forbidden from summarizing your contract back to you and calling it standard.
Fourteen prompts, in planning order
They run in the order the decisions actually arrive: split the total into a category budget, have the family money conversation before you tour anything, shortlist venues, get their real numbers in writing, decode and compare the quotes, price the guest list, read the contract, and build the day-of timeline. The dress, the flowers, the shower, the reception ideas, and the bachelor and bachelorette trip each have their own.
Every prompt shows the date we last checked it against the current models. When one stops working, we fix it or pull it.
Start where your next decision is: weddingish.com/ai.
