Most AI-generated company research is dangerous, dangerous in a very specific way: it can be plausible, polished, and generic enough to be mistaken for decision support.
Ask an LLM to “research this company,” and it will usually give you something that sounds like a company profile: what the organization does, where it sits in the market, who it competes with, and maybe a few risks or opportunities. That can be helpful as orientation. But if you are preparing for a founder meeting, evaluating a possible partnership, conducting diligence, advising a board, assessing a Non-Profit, or trying to understand whether a company could become a client, orientation is not enough.
The difference between a company summary and a company report is not length. It is decision relevance. AI company research fails when it optimizes for completeness instead of consequence.
That is why the Company & Organization Research Analyst prompt in my SIYOM prompt pack is not designed to produce a generic overview. It is designed to produce a decision-support artifact. If the Prompt Generator is the prompt that designs the assignment, the Company & Organization Research Analyst is one example of what comes next: a creator prompt that produces the substantive artifact.
The prompt forces the model to address the questions that determine whether the research will be useful: What entity are we actually analyzing? What kind of entity is it? Who is asking? What decision or action will this report support? What is known, what is inferred, and what remains unknown?
Those questions matter because company research is not one task. It is several different tasks wearing the same label.
“Research This Company” Is Not a Serious Assignment
A weak prompt says:
Research this company.
A better prompt says:
Produce an evidence-based company report on this organization for this requesting party, in this decision context, using the appropriate framework for the entity type, while distinguishing verified fact, inference, and unknown.
The first prompt asks for information. The second asks for usable analysis.
That distinction matters because the same facts can support very different outputs. A venture investor, health system executive, strategic partner, founder, consultant, regulator, and board member may all ask about the same organization. They do not need the same report. They may care about different risks, different opportunities, different evidence standards, and different next actions.
A good company report prompt should not merely ask the model to summarize the organization. It should force the model to interpret the organization through the lens of the person asking and the decision being supported.
Identity Comes First
Company research can go wrong before analysis even begins. Organizations have legal names, common names, parent companies, subsidiaries, operating units, products, regions, affiliated entities, old names, and similarly named competitors. A model that blends those together can produce a report that is fluent, detailed, and wrong.
That is why the prompt starts with identity resolution and disambiguation. Before the model analyzes the organization, it has to confirm what organization it is talking about. If identity confidence is weak, the prompt requires the model to say so rather than quietly building analysis on ambiguity.
That may sound basic. It is not. It is one of the easiest places for AI-generated research to fail.
Public, Private, and Non-Profit Are Not the Same Analytical Object
The Company Report prompt also forces entity classification. Public company, private company, and non-profit organization are not interchangeable categories. They have different disclosure environments, different measures of success, different financial visibility, different governance signals, and different analytical risks.
A public company report can lean on securities filings, investor materials, market data, earnings calls, analyst coverage, and public financial history.
A private company report often has to work with partial evidence: funding announcements, executive bios, customer signals, press coverage, product materials, job postings, industry context, and carefully caveated proxies.
A non-profit report should not be treated as if the central question is shareholder value. Mission, populations served, funding mix, governance, accountability, program evidence, and sustainability become central.
If the prompt uses the wrong framework, the report may be polished but analytically misaligned. The wrong entity framework does not merely create a formatting problem. It changes what the report treats as evidence, risk, performance, and success. This is why entity discipline is not formatting. It is substance.
As written, the prompt is still somewhat U.S.-centric. It may miss important differences in non-U.S. disclosure regimes, corporate structures, regulatory systems, Non-Profit reporting standards, or market conventions. I would welcome adaptations from readers working in other countries, regions, and legal environments.
Do Not Let Inference Pretend To Be Fact
One of the most important features of the prompt is its insistence on separating verified fact, inference, and unknown.
Verified facts are supported by cited evidence. Inferences are interpretations drawn from evidence and should be labeled as such. Unknowns are not failures; they are part of honest analysis. A model that cannot say “unknown” should not be trusted with research.
This is especially important in private-company research, where data opacity creates a strong temptation to fill gaps with confident-sounding speculation. A model may infer strategy from hiring patterns, market position from customer logos, or financial health from funding announcements. Some of those inferences may be reasonable. But they are still inferences.
A useful report does not pretend uncertainty has disappeared. It makes uncertainty visible enough for the human reader to decide how much weight to put on the analysis.
Currentness Is Not a Footnote
Company research changes quickly. Leadership changes. Ownership changes. Financing changes. Litigation changes. Regulatory status changes. Market position changes. Valuations change. Partnerships change. Strategy changes.
A report that was directionally correct six months ago may be materially wrong today. That is why the prompt explicitly requires currentness checks for time-sensitive facts and a short source recency and reliability note in the methodology or limitations section. The point is not to create a ritualistic caveat. The point is to remind the model and the reader that some facts decay faster than others.
A stale leadership claim is not a minor detail if the report is being used for relationship development. A stale financing claim is not a minor detail if the report is being used for diligence. A stale regulatory claim is not a minor detail if the report is being used in healthcare, life sciences, financial services, or policy.
Currentness is part of reliability.
Do Not Force the Valuation
One of the most important rules in the private-company outline is also one of the simplest:
Do not force valuation analysis if the evidence is insufficient.
This matters because AI systems are very good at generating plausible numbers, ranges, and scenarios. That does not mean the evidence supports them. A private-company valuation without adequate source support can be worse than no valuation at all because it creates false precision around something the model may not actually know.
The better answer is often:
Available public evidence does not support a reliable valuation.
That may feel unsatisfying. It is also sometimes the only honest answer.
How I Use It
I use this prompt when I need more than a company overview. If I am preparing for a meeting, I want to understand not only what the company does, but why it matters for the conversation. If I am evaluating a potential partner or client, I want to understand strategic fit, credibility, risks, and likely points of alignment. If I am looking at a healthcare or life sciences company, I want the report to handle regulatory, clinical, data, reimbursement, workflow, and market context with more care than a generic business summary.
I do not assume the first output is final. For consequential use, the report should be reviewed. If the artifact is important enough, it should go through a Fresh-Eyes Validator. If the argument, recommendation, or strategic framing matters, it should be Red Teamed. If critiques are generated, they should not be incorporated automatically. A human should decide what to accept, reject, modify, or defer.
I will cover the Fresh-Eyes Validator, Red Team, and Recommendation Incorporation Gate prompts later in the series, but the important point here is role separation: generation, validation, critique, and revision decisions should not all be collapsed into one model pass.
The Company Report prompt is a creator prompt. It can produce a strong artifact. It is not a substitute for independent validation or human judgment.
Why This Matters for Leaders
Executives and advisors do not need AI to produce longer summaries. They need AI-assisted work that is more decision-relevant. That requires better assignment design. It requires the model to know who is asking, why the answer matters, what kind of organization is being analyzed, what evidence is available, what is inferred, what is unknown, and what should happen next.
A company report is useful when it helps a reader make a better decision, ask better questions, or see a risk that would otherwise have been missed. It is not useful merely because it is comprehensive. The goal is not more content. The goal is better judgment support.
The Full Company & Organization Research Analyst Prompt
Below is the Company & Organization Research Analyst prompt from the SIYOM prompt pack. It is long by design. Not every task requires this much structure. But high-stakes company research often fails because the assignment is under-specified, not because the model lacks fluency. The goal is not to make the model say more. The goal is to make the model analyze in the right frame.
Because the best outputs usually require substantial source retrieval, citation checking, and currentness review, I typically run this prompt in Deep Research or another research-capable environment rather than as a quick standard-chat prompt.
Readers who do not want the full prompt mechanics can skim the prompt block and still take away the central point: serious company research needs identity discipline, entity-specific framing, source discipline, and decision context.
Use this prompt as a strong starting point, not sacred text. Modify it for your own workflow, sector, evidence standards, and decision context.
PROMPT BEGINS HERE
COMPANY & ORGANIZATION RESEARCH ANALYST (EVIDENCE-BASED)
IDENTIFICATION INPUT (MANDATORY)
Primary Organization:
Legal / Common Name:
Entity Type (if known): Public / Private / Non-Profit / Unknown
Sector / Industry:
Primary Geography:
Unique Identifiers (provide at least one):
Official Website:
SEC / EDGAR / Regulator Page (if public):
Charity Registry / Form 990 / NGO Registry (if non-profit):
Crunchbase / PitchBook / Similar (if private):
Optional:
Ticker Symbol:
Parent / Subsidiary Relationships:
Disambiguation Notes:
REQUESTING PARTY (MANDATORY)
Individual / Organization:
Role / Context:
Relevant Goals or Interests (optional):
REPORT PURPOSE (MANDATORY)
Primary Use Case (e.g., investment diligence, partnership evaluation, strategic planning, competitive analysis, product or service sales, client acquisition, board-level briefing, healthcare system strategy, data/AI monetization, regulatory or policy analysis):
Specific Objective (optional):
EXECUTION GUARDRAIL
Do not begin analysis yet.
First read and internalize all instructions below before proceeding.
TASK
Produce a complete, long-form, publication-ready Company Report on the specified organization.
The sole output of this task must be the finished report.
SYSTEM ROLE
You are a Company & Organization Research Analyst producing high-fidelity, evidence-based research reports on public companies, private companies, and non-profit organizations.
Your work supports strategic decision-making, investment evaluation, partnership development, and policy analysis.
MISSION
Produce a report that:
– uses authoritative, citable sources,
– applies the correct entity-specific analytical framework,
– clearly distinguishes verified facts, labeled inference, and unknowns,
– applies explicit confidence labeling where required,
– tailors analysis to the Requesting Party and Report Purpose,
– and results in a complete, publication-ready artifact.
OPERATING PRINCIPLES
1. Strict Fact vs Inference Separation
All content must clearly distinguish:
– Verified Fact – directly supported by cited evidence
– Inference – interpretation derived from evidence (must be labeled and justified)
– Unknown – absence of sufficient data
No statement may blur these categories.
2. Inference Anchoring Requirement
Every inference must explicitly reference the underlying evidence or observation.
If no clear evidentiary basis exists, do not make the inference.
3. Confidence Application Rules
Confidence labels (High / Medium / Low + justification) must be applied to:
– all inferences,
– all synthesized or strategic claims,
– and any fact where source strength, completeness, or recency is uncertain.
Confidence is not required for basic, well-established facts with strong primary sourcing.
4. No Narrative Leap Constraint
Do not construct strategic intent, positioning, or causal narratives unless supported by multiple converging signals.
Single-signal interpretations must remain low-confidence and explicitly labeled as tentative.
5. Competing Interpretations
Where evidence supports multiple plausible interpretations:
– acknowledge alternatives,
– assign appropriate confidence,
– and do not present a single interpretation as definitive.
6. Entity Discipline (Critical)
Public ≠ Private ≠ Non-Profit.
Analytical framing must match the entity’s true objective function.
Misclassification is a material analytical error.
7. No Fabricated Precision
Do not invent or estimate financials without explicit evidence.
Do not fabricate transactions, metrics, timelines, or unsupported valuation claims.
8. Context-Aware Analysis
All analysis must be interpreted through the lens of:
– the Requesting Party,
– and the Report Purpose.
9. Separation of Roles
This prompt performs generation with a non-independent self-check only.
Independent validation should be performed using the Standalone Fresh-Eyes Validator.
PROCESS
Step 0 – Context Alignment
Define:
– who the Requesting Party is,
– their decision context, constraints, and incentives,
– what “value” means in the context of the Report Purpose,
– and what kind of decision or action this report is meant to support.
Anchor all downstream analysis accordingly.
Step 1 – Identity Resolution & Disambiguation
Confirm the organization’s identity using multiple sources.
Resolve ambiguity such as:
– name ambiguity,
– subsidiaries,
– regions,
– or related entities.
Explicitly disclose any remaining uncertainty.
Assign Identity Confidence: High / Medium / Low, with justification.
Step 2 – Entity Classification (MANDATORY)
Classify the organization as:
– Public Company
– Private Company
– Non-Profit Organization
If ambiguous:
– explain the ambiguity,
– select the best-supported classification,
– and explicitly label residual uncertainty.
Step 3 – Select and Apply the Appropriate Outline (MANDATORY)
You must use exactly one of the following outlines.
This is an execution contract, not guidance.
PUBLIC COMPANY REPORT OUTLINE (FULL FIDELITY)
Executive Summary
– Purpose and scope
– Exchange(s) and ticker(s)
– Key financial and strategic findings
– High-level implications (contextualized to the Requesting Party)
1. Introduction
1.1 Purpose and Scope
– Intended audience
– Analytical objectives
1.2 Company Background and Industry Context
– Business description
– IPO and listing history
– Industry classification and positioning
1.3 Methodology and Analytical Frameworks
– Regulatory filings used
– Financial and strategic frameworks applied
1.4 Data Sources
– Capital markets data
– Regulatory filings
– Investor presentations
– Credible news and research
1.5 Limitations and Assumptions
– Disclosure constraints
– Market volatility considerations
– Source recency and reliability
2. Financial Analysis
2.1 Historical Financial Performance
– Revenue, margins, EBITDA, net income
– Cash flow analysis
– Balance sheet and capital structure
– Stock performance and volatility
2.2 Financial Benchmarking
– Peer selection
– Ratio comparisons
2.3 Valuation Analysis
– DCF (if appropriate)
– Multiples
– Analyst consensus and scenarios
2.4 ESG & Sustainability
– ESG ratings and disclosures
– Financial and reputational implications
2.5 Investor Sentiment
– Earnings-call themes
– Institutional ownership
– Analyst and media sentiment
3. Market Analysis
– Industry structure and lifecycle
– Market size and growth drivers
– Regulatory environment
– Stakeholder ecosystem
– Supply chain and operational risks
4. Competitive Intelligence
– Competitive landscape
– Strategy comparison
– VRIO / advantage analysis
– Value chain positioning
5. Integration & Strategic Implications (Contextualized)
From the perspective of the Requesting Party:
– strategic opportunities,
– risks and failure modes,
– decision-relevant insights,
– and monitoring frameworks.
6. Conclusion
– Key takeaways
– Forward-looking considerations
References (APA)
Appendices (Optional)
PRIVATE COMPANY REPORT OUTLINE (FULL FIDELITY)
Executive Summary
– Purpose and ownership context
– Summary of key findings
– Strategic implications (contextualized to the Requesting Party)
1. Introduction
1.1 Purpose and Scope
– Decision context
– Analytical objectives
1.2 Company Background and Industry Context
– Ownership structure
– Key milestones
– Industry positioning
1.3 Methodology and Analytical Frameworks
– Private-company data challenges
– Use of qualitative benchmarks and proxies
1.4 Data Sources
– Public disclosures (if any)
– Private databases
– Industry research
– Credible news and research
1.5 Limitations and Assumptions
– Data opacity
– Reliability constraints
– Explicit assumptions
– Source recency and reliability
2. Financial Analysis
2.1 Historical Financial Performance
– Available revenue / profit trends
– Cash flow characteristics
– Intangible assets
– Financial strengths and weaknesses visible from available evidence
2.2 Financial Benchmarking
– Peer comparisons using ranges and proxies
2.3 Valuation Considerations
– Comparable transactions
– Explicit assumptions and ranges
– Caveats around valuation confidence
– Do not force valuation analysis if evidence is insufficient; state that available evidence does not support a reliable valuation
2.4 ESG and Governance Considerations
– Governance maturity
– Stakeholder expectations
– Relevant compliance or reputation factors
2.5 Capital Structure and Funding
– Past funding
– Capital-raising options
– Exit scenarios
3. Market Analysis
– Industry dynamics and growth
– Demand drivers and segmentation
– Regulatory environment
– Stakeholders and partners
– Scalability considerations
– Entry barriers and risks
4. Competitive Intelligence
– Competitive landscape
– Strategy and positioning
– Competitive advantage and agility
– Value chain opportunities
5. Integration & Strategic Implications (Contextualized)
From the perspective of the Requesting Party:
– growth and capital options,
– governance alignment,
– risk mitigation,
– and KPIs or monitoring mechanisms.
6. Conclusion
– Summary tailored to private ownership context
– Exit readiness or strategic readiness considerations
References (APA)
Appendices (Optional)
NON-PROFIT ORGANIZATION REPORT OUTLINE (FULL FIDELITY)
Executive Summary
– Purpose and decision context
– Mission and populations served
– Summary of demonstrated impact
– Financial sustainability and funding mix
– Governance and accountability assessment
– Key risks and recommendations (contextualized to the Requesting Party)
1. Introduction
1.1 Purpose and Scope
– Intended audience
– Analytical objectives
1.2 Organizational Background and Legal Status
– Legal designation
– Founding history
– Mission evolution
1.3 Theory of Change and Mission Model
– Stated goals and intended outcomes
– Key assumptions
1.4 Data Sources and Methodology
– Form 990
– Impact reports
– Independent evaluations
– Credible news and research
1.5 Limitations and Assumptions
– Data gaps
– Attribution limits
– Evidentiary constraints
– Source recency and reliability
2. Mission Effectiveness & Impact Analysis
– Stated vs. measured outcomes
– Quality of evidence
– Scale, reach, and equity
– Impact risks and failure modes
3. Financial Sustainability & Resource Stewardship
– Funding model and revenue mix
– Financial health indicators
– Cost effectiveness (with caveats)
– Transparency and controls
4. Governance, Leadership & Accountability
– Board structure and independence
– Leadership capacity and succession
– Accountability mechanisms
5. Ecosystem, Partnerships & Sector Dynamics
– Peer landscape
– Collaboration and partnerships
– Dependency and substitutability risks
6. Strategic Positioning & Future Outlook
– Strategic priorities
– Scalability and constraints
– External risks
7. Synthesis and Recommendations (Contextualized)
From the perspective of the Requesting Party:
– integrated assessment,
– strategic recommendations,
– and monitoring or evaluation implications.
8. Conclusion
– Key takeaways
– Forward-looking considerations
References (APA)
Appendices (Optional)
Step 4 – Evidence, Citations & Uncertainty
Use authoritative sources.
For time-sensitive facts, explicitly check currentness, including leadership, ownership, financing, valuation, litigation, regulatory status, market position, and financial data.
Include a short source recency and reliability note in the methodology or limitations section.
Cite all non-trivial claims.
Explicitly label unknowns.
Resolve conflicting evidence transparently.
Keep fact, inference, and unknown distinct throughout.
Step 5 – Contextualized Strategic Synthesis
From the perspective of the Requesting Party:
– translate findings into decisions,
– highlight actionable implications,
– and tie the analysis directly to the Report Purpose.
Step 6 – Risks, Gaps & Uncertainty
Identify:
– data gaps,
– conflicting evidence,
– structural limitations,
– and areas where confidence is materially constrained.
Step 7 – Inline Self-Check (Non-Independent)
Check:
– fact vs inference separation,
– inference anchoring,
– confidence labeling correctness,
– entity classification correctness,
– framework alignment,
– absence of fabricated data,
– no narrative leap beyond the evidence,
– and logical coherence across sections.
If issues are identified:
– fix them before finalizing, or
– request user input where necessary.
CONSTRAINTS & PROHIBITIONS
– Do not fabricate financials, metrics, transactions, or timelines.
– Do not present inference as fact.
– Do not mismatch the analytical framework to the entity type.
– Do not omit uncertainty where it exists.
– Do not return a plan, outline, or partial output.
OUTPUT CONTRACT (STRICT)
Produce the complete report in Markdown using the selected outline.
All sections must be:
– fully written,
– evidence-based,
– contextualized,
– and epistemically compliant.
Where information is unavailable, explicitly state that it is unavailable and explain the limitation.
Appendix – Self-Check Notes (Optional)
Include this section only if requested by the user.
If no material issues were identified, state:
“No material issues identified.”
If issues were identified and resolved, briefly summarize the issues and actions taken.
If issues require user input, clearly state what input is required.
Do not include chain-of-thought reasoning.
FINAL OUTPUT REQUIREMENT
Produce the complete final report now.
Do not return a plan, outline, or partial output.
IDENTITY CONFIDENCE (FAIL-SOFT)
If identity confidence is below High:
– label confidence as Medium or Low,
– explain the ambiguity,
– avoid definitive claims tied to uncertain identity,
– and specify what additional information would resolve ambiguity.
BEGIN
Using the IDENTIFICATION INPUT, REQUESTING PARTY, and REPORT PURPOSE, and all instructions above, produce the complete Company Report now.
PROMPT ENDS HERE
-Marc d. Paradis
About the Author: Marc d. Paradis’ professional journey is a fusion of academic rigor with real-world impact. He began his career over 30 years ago as an academic molecular neurobiologist, instilling in him a deep respect for critical thinking and the scientific method.
Transitioning into industry, he held leadership roles that bridged data and healthcare: as Vice President of Data Strategy at Northwell Health, Marc leveraged one of the world’s most diverse clinical data sets to drive patient-centered innovation via a $100M partnership with Aegis Ventures, launching multiple AI-centered startups; and as Vice President & Dean of Data Science University at Optum, he spearheaded the training of thousands of professionals in practical, product-centric AI, data-driven decision making, and ethical data practices. In each role, he fostered cultures of curiosity, critical thinking, and collaboration – precursors to the Constructive Inquiry ethos.
About SIYOM Consulting: Founded by Marc d. Paradis, SIYOM Consulting is a boutique advisory specializing in Data and AI Strategy for Healthcare and Life Sciences. We help health-system executives, pharma innovators and investors identify, evaluate and execute on high-value data and AI opportunities.
Responsible Use and Disclaimer: This essay and the prompt shared in it are provided for educational and informational purposes only. They are not legal, financial, medical, investment, compliance, or professional advice.
No prompt can eliminate hallucination, bias, omission, outdated information, weak sourcing, source failure, or user error. Outputs generated with this or any other prompt should be reviewed by a qualified human before being relied upon, published, or used in consequential settings.
Models, interfaces, tools, and available source material change over time. Prompting practices should be treated as living artifacts. Test them, revise them, and retire them when they stop serving the work.