# Reasons to Invest Versus the Sources ## An external investor team made six statements in favor of the investment The sources do not support all of them. An external investor team looked at the company. The team sent a summary of its positive points. This page compares each point with the other pages in this map. The summary is an opinion. It is not a result. | Statement of the investor team | What the sources show | Status | | --- | --- | --- | | Loss of tribal knowledge is a known problem in the industry. Many factories in the target region use systems that are not new. The work depends on technicians with the know-how to find faults. | Research that has no relation to the company documents the pain. No source shows that customers pay for a solution at a large scale. See [[The Tribal Knowledge Thesis]]. | PARTIAL | | The software is not only an AI wrapper. The language model decides how to do complex tasks. A person does not specify the workflow. | The architecture has good parts and a step for human approval. No source tests how much the model does by itself. See [[The Four Layer Stack and What Is Commodity]]. | CLAIMED | | A multimodal retrieval system works with no knowledge cutoff. | Many systems have retrieval. Retrieval does not keep competitors out. See [[The Four Layer Stack and What Is Commodity]]. | CLAIMED | | Exclusive partnerships with institutes and access to proprietary data make the software hard to copy. | No contract in the sources shows exclusivity. Other parties have the data rights. See [[Borrowed Data Is Not a Moat]]. | UNSUPPORTED | | The founders have experience and technical skills. | The founder is the strongest asset. The sources do not show the engineering team. See [[Team Depth and Key-Person Concentration]]. | PARTIAL | | The team recently came in the top ten of a start-up challenge of a GPU vendor. More than 500 start-ups took part. | Only the investor team says this. No source shows it. A challenge selects teams. It does not validate the technology. | CLAIMED | > [!important] The correct test > The sources agree that the company is not a thin wrapper. The correct test has two parts. First, the deployments must share the platform layer. Second, the company must have the data rights. See [[Platform or Solutions - The Central Question]]. > [!tip] A method for technical judgment > The investor team wants a technical view of how complex and how different the software is. This map gives a method with three steps. > Step 1 divides the parts into those that all vendors have and those that are different. Step 2 tests how much code the deployments share. Step 3 tests the data rights and the IP. > See [[The Four Layer Stack and What Is Commodity]], [[Delivery Economics and the Scaling Ceiling]] and [[IP Encumbrance Map]]. ## Why it matters The investor team lists what the company says about itself. A diligence map lists what the sources support. The thesis and the founder are the strongest points. The data moat and the autonomy of the model are the weakest points. ## Related - [[Foreman MOC]] - [[Claims Versus Evidence]] - [[What Must Be True]]