Company AGI

A concept describing the trajectory of enterprise AI development, in which every company is effectively building its own internal form of artificial general intelligence — an integrated AI system capable of handling the full scope of that company's operations, decisions, and customer interactions.

Definition and framing

The term surfaces in the context of a broader framework for how organizations should think about AI investment. The speaker in The CEO Must Be the Chief AI Officer endorses what they call the "Jack Dorsey view": that every company is essentially trying to build its own little Company AGI. 1:12:37

Component structure

Company AGI is treated as an umbrella concept composed of three distinct layers, each a part-of the whole: 1:08:28

• Product AI — AI embedded in or delivered through the company's external products.
• Operational AI — AI that directly affects the company's ability to serve customers at scale. 1:08:34
• Corporate AI — AI governing how people work internally. 1:08:43

A Customer World Model is also described as a constituent component — a dedicated agent that models the company's customers and their problems, on top of which further agents (such as a product roadmap manager) can be built. 1:14:06

Points of disagreement

While the speaker accepts the goal of Company AGI, they explicitly reject a naive implementation path. The speaker criticizes the idea of a single monolithic company model that ingests every piece of data without judgment or lens: "I don't believe in the like oh I'm going to have like a single company model that has like every piece of data like in a single like with no judgment or lens into anything." 1:13:21

Instead, the preferred architecture decomposes Company AI into domain-specific agents with clear boundaries — for example, a dedicated customer-world agent and a separate product-roadmap agent built on top of it — rather than collapsing everything into one undifferentiated model. 1:13:21 1:14:06