How to Reorg After AI Changes Everything | Block's Owen Jennings on the a16z Show

Block's AI-Driven Restructuring and Workforce Reduction

• Workforce Reduction: Block executed a reduction in force (RIF) of slightly greater than 40% of its total workforce.
• Targeted Cuts: The cuts were heavily concentrated on the engineering and product teams, while outbound sales and account management roles saw negligible reductions ("de minimis").
• Strategic Timeline: The decision followed a two-year period of internal preparation, including the early 2024 launch of "Goose," an agent harness, and intensified evaluation following late 2024 releases of Opus 4.6 and Codex 5.3.
• Productivity Shift: Executives state the historical correlation between headcount and output "broke" in early December, with 1-2 engineers utilizing AI tools achieving 10x to 100x productivity gains.
• Organizational Design: The company restructured from a hierarchical, functional model to small "squads" of one to six people with significantly reduced management layers (estimated 50-60% reduction in development layers).
• Operational Principles: The transition prioritized three non-negotiables: maintaining P0-level reliability, preserving the compliance and regulatory team intact, and continuing long-term roadmap features despite smaller team sizes.

AI Infrastructure and Operational Automation

• Internal Tooling: Block deployed "Goose," a model-agnostic agent harness supporting over 120 models, and "G2," an internal agentic operating system for automating deterministic workflows.
• Software Development: The "Builder Bot" autonomously merges pull requests and builds features, typically reaching 85-90% completion before human review of the final 10%.
• Employee Workflow: Engineers, designers, and product managers now ship PRs directly; the workflow has shifted from linear sequential processing to managing multiple parallel agents (e.g., 14 agents building PRs simultaneously).
• Support Automation: AI and chatbots now automate the majority of customer support inquiries, including voice and text channels, shifting human roles to "human-in-the-loop" oversight.
• Generative UI: Block is transitioning from static user interfaces to generative UI, where visualizations and app features (e.g., MoneyBot charts, ManagerBot scheduling apps) are dynamically generated based on user prompts rather than hardcoded into the source code.

Strategic Outlook and Competitive Moats

• Product Consolidation: The company moved away from siloed business units (Square, Cash App, Afterpay) to a unified functional structure, building shared financial and business platforms across all product lines.
• Jevons Paradox: While fewer engineers are needed per product, Block anticipates a net increase in tech companies and development activity across previously inaccessible sectors due to lowered barriers to entry.
• Defensibility: The primary long-term competitive moat is defined as the depth of a company's proprietary understanding of specific hard-to-understand domain data, rather than just code or distribution.
• Risk Assessment: Executives acknowledge the "nightmare" QA challenges of generating non-deterministic outputs for tens of millions of users, necessitating robust internal world models and feedback loops.
• Cultural Impact: The transition involved an aggressive but transparent communication strategy, including all-hands meetings with leadership, generous severance, and a 70-80% reduction in internal meetings to prioritize building time.
• Future Vision: The organization is building toward an "intelligent system" where the company iterates on its own operating model and customer insights hundreds or thousands of times daily via agentic systems.