Defining the Uncharted Territory of Business Development × AI as 'Deep Out'

While AI-assisted coding has already established a market, there are currently no documented or systematized methodologies for bringing AI collaboration into business development—the intellectual work that spans organizations, customers, finance, and competition.

One executive, working with a $100/month AI editor, generated 393 ideas over 145 hours across 25 days. Nine of these were evaluated as patent-worthy, with two currently in the application process. This represents 32 times the productivity of standard operations—but the most significant discovery was not the numbers, but that the AI's behavior changed simply by deepening the context of the dialogue, without changing the model itself.

Approaches to improving AI output have historically focused on 'Scale Out'—developing higher-performance models. 'Deep Out' (Deepening Extension) is a different approach. By deepening the context between human and AI without changing the model, the quality of output shifts. This synergy allows individuals to create value that surpasses entire organizations.

How it began: In the summer of 2025, at a dining table in a Tokyo apartment, an executive began collaborating with AI. Junya Sato (Jay), who has worked at Recruit, Oracle, Google, and Salesforce, and currently runs an IT consulting firm, is neither an AI researcher nor an engineer. He used a $100/month AI editor. Unlike standard web-based AI, which requires copy-pasting and loses context, he used Claude Code (Anthropic), which can read and write files directly on a computer. By sharing business plans, competitive analysis, patent strategies, and financial models with the AI, he built his thinking hour by hour.

From 145 hours of intensive dialogue, 393 ideas were born. Nine were deemed patent-worthy. This was 32 times the density of standard business productivity. What emerged was not just 'faster work,' but 'structural insights unreachable alone' and a new 'density of creation.'

By not changing the model, not fine-tuning, but simply deepening the context of the dialogue, the AI's behavior changed. Sato has named this phenomenon 'In-Context Adaptation.'