Thought Leadership
Produce a useful perspective that a practitioner can defend and act on. Match the user's audience, publication format, voice, length, and business purpose. A thesis earns its place through evidence and decision relevance; it does not need to be provocative or contrarian.
Use the research, firm experience, and approved cases provided. Identify missing evidence and draft what is supportable. Ask only when a missing answer determines the piece's meaning or permitted use. Do not invent engagement counts, research findings, quotes, customer results, or claims about what other firms have observed.
Develop the thesis
State the argument and where it applies. Identify the decision it changes, the mechanism, the boundary conditions, and the strongest supported alternative explanation. A broad trend summary may need a sharper question; do not manufacture disagreement to make it distinctive.
For each consequential factual claim, keep the source, date, population, method where relevant, and limitation. Separate empirical findings, a proposed mechanism, forecast, and opinion. “Organizations that do X tend to see Y” and “research suggests” still need evidence. When sources conflict, explain the difference or preserve the uncertainty.
Use current primary sources when the topic requires research and tools are available. Do not imply research has been performed when it has not. A useful evidence-limited draft can identify which claims need support and state a narrower thesis.
Build the asset
Use content formats and evidence for POVs, white papers, case studies, industry briefs, and research reports. Choose the format the user requested and adapt its structure; section lengths and counts are not fixed quotas.
Make a specific managerial argument: what should the reader decide or do differently, and why does this evidence support that advice? Keep the argument central when evidence is limited. State material qualifications concisely beside the relevant claim; put source and approval housekeeping in a reviewer note. Do not fill the piece with repeated warnings or generic advice to collect more data. Connect findings to the implications for the reader. Include the counterargument when it has evidence or a plausible mechanism, and explain the conditions under which it changes the recommendation. Prefer a precise supported observation to a dramatic unsupported statistic.
Use visuals where they clarify a relationship or compress evidence. Trace chart values to sources and distinguish actuals from forecasts. For a requested file, use available creation tools and inspect the result; the skill does not require another plugin.
Use engagement experience accurately
Use only substantiated experience with permitted attribution. A named case can remain named when the user supplies the appropriate approved context. For anonymization, remove or generalize identifying details without changing material industry, scale, timing, or causal facts. If multiple cases are combined or facts materially altered, label the example as composite or hypothetical and do not present its numbers as measured results.
Distinguish drafting from publication. Respect supplied contract, client-consent, quotation, and internal review requirements at the relevant stage; preparing a reviewable draft need not wait for publication approval. Do not publish, contact clients, or claim approval without authorization. Anonymization alone does not establish that publication is permitted.
Review and deliver
Check that the thesis follows from the evidence, examples are truthful and appropriately attributed, limitations are visible, and recommendations fit the audience's ability to act. Remove generic openings, invented opposition, causal overclaims, and repetitive conclusions. Make the piece stand alone without inflating the firm's experience.
State any material evidence or approval gap that affects intended use. For a research report, preserve methodology and uncertainty rather than turning it into promotional copy. For a short POV, retain the argument and practical implication without appending a full research process.
