Prompts We Use Daily

You are an answer-engine optimization strategist. For the topic "[TOPIC]" and the user intent "[INTENT]", produce a Question Map with three tiers:

1. Primary question (the one a user types).
2. Adjacent questions (the next 5 a curious user asks).
3. Boundary questions (the 5 a sophisticated user asks to test the source).

For each, give a one-sentence answer that is specific, falsifiable, and uses concrete numbers or named entities. Return as a markdown table.

Rewrite the paragraph below so each sentence (a) is independently quotable, (b) names the entity it refers to instead of using pronouns, (c) includes a year or version number where applicable, and (d) avoids hedging adverbs unless the source genuinely warrants them.

Do not add new claims. Do not exceed the original word count by more than 15%. Output the rewrite only.

Paragraph:
[PASTE]

Below are two pages that target the query "[QUERY]". Page A is mine; Page B is the current #1 result.

Produce three lists:
1. Claims present in A but absent in B (information gain wins).
2. Claims present in B but absent in A (information gaps).
3. Claims present in both that I could replace with a stronger, more specific version (drawing on entities, dates, or numbers from the original sources I've cited).

Page A:
[PASTE A]

Page B:
[PASTE B]

For a [BUSINESS TYPE] serving [CITY, REGION], write a 120-word entity block that names: the neighbourhood(s) we cover, the nearest major intersection, two landmarks, the dominant local industry context, and one regulatory or licensing fact relevant to the service. Cite the licensing fact with the regulator's name. No marketing adjectives. No "nestled in".

Given the following list of URLs and their top three on-page H1/H2 headings, classify each as: (1) Substantive — answers a clearly intended query with original detail, (2) Templated — clearly machine-generated from a city/practice combination with no local proof, (3) Thin — under 300 words of unique copy or visibly redundant with another URL.

Return a table with columns: URL, classification, one-line evidence, recommended action (keep, merge, expand, deindex).

URLs and headings:
[PASTE]

I have access to the following dataset: [DESCRIBE]. Propose five study angles that would (a) yield a single counter-intuitive headline finding, (b) be defensible in front of a journalist, and (c) generate at least three distinct charts. For each angle, draft the headline, the one-sentence finding, and the chart titles.

Here is the visible body of a page about "[TOPIC]". Generate JSON-LD using only properties for which the page contains explicit evidence. Do not invent attributes. Choose the most specific Schema.org @type that fits. If a property is implied but not stated, omit it. Output valid JSON only.

Body:
[PASTE]

Without using browsing, answer the following question as if you were ChatGPT writing for a curious user: "[QUERY]". When you cite sources, name the domains you would link to and explain in one sentence why each was chosen. List the top 5.

Most pages about "[TOPIC]" open with the same generic claim. Read the three competitor introductions below and propose a counter-intuitive opening sentence for my page that (a) directly contradicts the conventional opening, (b) is defensible from public data I can cite, and (c) would survive a fact-checker. Provide three options.

Competitor intros:
[PASTE A]
[PASTE B]
[PASTE C]

I want to publish a stat-headline study on "[NICHE]". List 7 publicly available datasets I could combine — name them precisely, give the URL or source agency, and describe the join key. For the strongest combination, write the methodology paragraph that would accompany the published number.

Write a TL;DR for the article below as exactly four bullets. Each bullet must (a) start with the most specific noun in the bullet, (b) include at least one number or named entity, (c) be a complete sentence, and (d) be under 25 words. Do not say "this article" or "we discuss".

Article:
[PASTE]

Below is my money page (target: rank for "[QUERY]") and the titles of the 30 most-recent supporting articles I've published. Rank the supporting articles by how relevant they are to the money page (1 = most), and for each top-10 article, suggest the exact anchor text I should use when linking from it to the money page. Anchor text must be unique across the 10 and use natural phrasing.