Build an AI-powered research tool that evaluates how easy or difficult it would be for another...

Build an AI-powered research tool that evaluates how easy or difficult it would be for another company to replicate a product or business offering, using all relevant publicly available information about the company. Core objective Given a company name, website, product page, or other identifying information, the system should research the company and produce a structured Product Replication Difficulty Assessment. The tool should answer: "If a competent competitor had access to the information that is publicly available about this company, how difficult would it be for them to create a substantially similar product?" This is an analytical assessment, not a claim that the company can actually be copied without legal, technical, or commercial consequences. Research sources The agent should gather and cross-reference as much relevant public information as possible, including: Company website Product pages Pricing pages Product documentation Help centres FAQs Terms and conditions Privacy policies Press releases News articles Interviews Founder statements Investor presentations Public filings, where applicable Job advertisements LinkedIn/company information GitHub and other public repositories Public APIs and developer documentation App Store / Google Play listings Product reviews Customer reviews Reddit and other public communities Social media posts YouTube demonstrations Archived versions of the website Public technical documentation Patents or other publicly available intellectual-property information The system should distinguish between verified information, reasonable inference, and speculation. Replication dimensions Score the product across several independent dimensions. 1. Product complexity How technically and operationally difficult is the product to reproduce? Consider: Software complexity Hardware complexity Manufacturing requirements Infrastructure Data requirements Algorithms/models Integrations Real-time systems Scalability Reliability requirements 2. Proprietary technology Determine whether the company appears to possess technology that would be difficult to reproduce. Look for: Proprietary algorithms Proprietary software Proprietary models Custom infrastructure Patents Custom hardware Proprietary datasets Internal tooling Unique technical processes Do not assume something is proprietary merely because the company describes it as proprietary. Identify the evidence. 3. Data advantage Assess whether the product depends on data that a competitor would struggle to obtain. Consider: Proprietary datasets Historical data User-generated data Training data Network-generated data Data partnerships Data collection scale Data quality Data accumulated over time 4. Capital requirements Estimate the resources required to reproduce the product. Consider: Initial development costs Hardware Infrastructure Employees Manufacturing Regulatory costs Marketing Customer acquisition Working capital Provide a qualitative estimate where precise figures cannot be established. 5. Human expertise Assess whether reproducing the product requires unusually specialised expertise. Consider: Engineering expertise Scientific expertise Industry expertise AI/ML expertise Manufacturing expertise Regulatory knowledge Specialist personnel 6. Time to replicate Estimate how long a competent competitor might require to create: A basic functional prototype A commercially viable version A version comparable to the company's current product Clearly distinguish estimates from verified facts. 7. Infrastructure and distribution Assess whether replication requires access to difficult-to-obtain infrastructure. Examples: Supplier relationships Manufacturing capacity Cloud infrastructure Distribution networks Retail relationships Logistics Regulatory approvals Payment infrastructure Data partnerships 8. Network effects and ecosystem Determine whether the product becomes harder to replicate because of: Existing users Marketplace liquidity Community Developer ecosystem Integrations Partnerships Network effects Switching costs 9. Brand and customer trust Separate the product itself from advantages created by: Brand recognition Reputation Customer trust Existing customer relationships Reviews Community A product may be technically easy to replicate while the business surrounding it is difficult to replicate. 10. Legal and regulatory barriers Assess whether replication could be affected by: Patents Copyright Trademarks Licensing Regulatory approvals Industry-specific regulations Contractual restrictions Do not provide legal conclusions. Flag areas that would require professional legal review. Important distinction The system must produce TWO separate assessments: A. Product Replication Difficulty How difficult would it be to reproduce the actual product/service? B. Business Replication Difficulty How difficult would it be to reproduce the broader business advantage, including customers, distribution, brand, data, partnerships, and infrastructure? This distinction is essential. For example: A product might receive: Product replication difficulty: Low Business replication difficulty: High because the underlying technology is relatively straightforward but the company has a large customer base, proprietary data, partnerships, and distribution network. Scoring system Create a transparent scoring model from 0 to 100. Do not simply generate a number using an LLM impression. Calculate the score from the underlying dimensions. For example: Technical complexity: 15% Proprietary technology: 15% Data advantage: 15% Capital requirements: 10% Human expertise: 10% Time to replicate: 10% Infrastructure/distribution: 10% Network/ecosystem effects: 5% Brand/customer trust: 5% Legal/regulatory barriers: 5% Show the individual component scores and explain how each was determined. Evidence requirement Every significant assessment should have supporting evidence. For each dimension, show: Finding Evidence Source Confidence level Impact on replication difficulty Use confidence levels: High Medium Low The system should never present an inference as an established fact. Competitor simulation After completing the research, create a hypothetical scenario: "Assume a well-funded competitor has competent engineers, access to normal commercial suppliers and access to all publicly available information identified by this analysis." Then estimate what they could realistically reproduce. Break this into: 30 days 6 months 1 year 3 years For each period, identify what could plausibly be reproduced and what would remain difficult. Final report The final report should contain: Company overview Product being assessed Executive summary Product replication score Business replication score Dimension-by-dimension analysis Evidence and sources Proprietary technology analysis Data moat analysis Infrastructure analysis Capital requirements Estimated replication timeline Legal/regulatory considerations What a competitor could reproduce immediately What would take significant investment What would be difficult to reproduce Key uncertainties Source list Anti-hallucination requirements The system must: Never invent company technology Never invent financial information Never assume a technology is proprietary without evidence Clearly label estimates Clearly distinguish facts from inference Cite sources for important claims Prefer primary sources Cross-check important claims against multiple sources State when information cannot be found Avoid treating marketing claims as independently verified facts Output Create a clean interface where the user can enter: Company name Company URL Product URL Optional additional information Then show the research progress and ultimately produce an interactive report. Include: Overall replication difficulty Product vs business replication comparison Dimension scores Evidence Sources Confidence levels Timeline Key barriers Key vulnerabilities/advantages The underlying research and scoring methodology should be visible to the user so that the result is auditable rather than being a black-box AI score.