Useful prompts specify context, evidence, constraints, output structure and measurement. These examples are intentionally practical: replace the business context with your own data and require the model to flag missing evidence instead of fabricating it.
SEO
Content refresh
Audit an existing page against current search intent. Separate evidence from assumptions, identify information gaps, recommend sections to keep/update/remove, and define a measurement plan for rankings, qualified clicks and conversion.
Marketing
Campaign brief
Turn a business goal, audience, offer and constraint set into a campaign brief with one primary message, supporting proof, channel roles, creative hypotheses, KPIs and a 14-day learning agenda.
Sales
Discovery call
Prepare a discovery plan that tests urgency, business impact, decision process, existing alternatives and cost of inaction. Produce questions, follow-ups, disqualification signals and a concise post-call summary template.
Analytics
Funnel diagnosis
Diagnose a conversion funnel from supplied stage volumes. Quantify drop-offs, identify the two highest-leverage bottlenecks, propose experiments, define guardrail metrics and state which conclusions cannot be made from the data.
Product
PRD outline
Create a decision-ready PRD from the problem statement. Include user problem, evidence, non-goals, scope, acceptance criteria, instrumentation, launch risks and unresolved questions before proposing features.
Operations
SOP
Convert a repeated operational task into an SOP with trigger, owner, inputs, step-by-step procedure, quality checks, exception handling, evidence retention and review cadence.
E-commerce
Product page
Build a product-page messaging hierarchy using audience problem, product mechanism, proof and objections. Produce headline options, benefit blocks, FAQ, trust elements and test hypotheses without inventing claims.
Strategy
Scenario planning
Create bull, base and bear operating scenarios from explicit assumptions. Show the causal drivers, leading indicators, decision thresholds, mitigation actions and what evidence would make each scenario more or less likely.