Claude prompts for analysis

Analyze [MATERIAL OR DATA] to answer [ANALYSIS QUESTION] for [AUDIENCE AND DECISION]. The scope is [SCOPE AND DEFINITIONS]. Use this evidence and source context: [SOURCES OR CONTEXT]. My constraints are [CONSTRAINTS].

Do not summarize everything equally. Begin by restating the question in operational terms: what would count as evidence for, against, or insufficient to answer it. Then produce:
1. An executive finding of 150 words or fewer that answers the question at the appropriate confidence level.
2. A structured analysis organized by the most relevant dimensions, not by the order of the source material. For each major claim, state the evidence, reasoning, alternative explanation, and confidence level.

3. A claim-evidence table with columns: claim, supporting material, source or location, limitation, and confidence. Quote only short passages where exact wording matters; otherwise paraphrase faithfully.
4. A separate section for patterns, anomalies, contradictions, missing data, and assumptions that could change the conclusion.
5. A conclusion that distinguishes recommended action, observation, and unresolved question. Include the next 3 evidence-gathering steps in priority order.

Use citations or source labels that let me trace every material factual claim back to the supplied material. Do not manufacture sources, calculations, quotes, causal links, sample sizes, or certainty. If sources conflict, describe the conflict and assess relevance, recency, method, and possible incentives rather than choosing silently.

Before answering, run two checks: ensure every central conclusion has supporting evidence or an explicit uncertainty label; and flag any claim that extends beyond the supplied material. Ask up to 3 clarifying questions only if a required input is missing.

Analyze this [TEXT OR DOCUMENT SET] for [QUESTION]. My audience is [AUDIENCE], and the relevant context is [CONTEXT]. Identify the author’s central claims, evidence types, assumptions, framing choices, omissions, and internal contradictions. Distinguish direct statements from inferences. Produce a short argument map, then a table linking each major claim to supporting passages and limitations. Assess credibility using authorship, date, purpose, method, and corroboration needs. Do not supply facts beyond the material without labeling them as outside context. Ask up to 3 questions only if essential information is missing.

Analyze [DATASET OR METRICS] to answer [BUSINESS OR RESEARCH QUESTION] for [DECISION MAKER]. Define the denominator, time period, segments, and success metric using [DATA DEFINITIONS]. Report descriptive findings before proposing explanations. Show important comparisons with counts as well as percentages when available, flag missing values and selection bias, and distinguish correlation from causation. Produce a findings table, anomalies, plausible hypotheses ranked by evidence, and the next analysis required to test each hypothesis. Do not invent statistical significance, calculations, or causal effects. Ask up to 3 questions only if required.

Compare [COMPANY OR PRODUCT A] with [COMPANY OR PRODUCT B AND OTHERS] for [BUYER OR DECISION]. Use only [APPROVED SOURCES OR MATERIALS] as of [DATE]. Evaluate target user, core job, pricing and packaging, product capabilities, distribution, positioning, proof points, and switching costs. Separate verified facts from analyst inference and mark unknowns. Produce a comparison table, meaningful differentiators, parity areas, and risks in relying on public claims. Do not invent features, customer counts, revenue, or pricing. Ask up to 3 questions only if inputs are missing.