How Block’s AI Agent ‘Goose’ is Revolutionizing Engineering and Beyond

In a development that blurs the lines between current reality and the perceived AI future, engineers at Block are leveraging a sophisticated AI agent to dramatically accelerate their workflow. This isn’t a hypothetical scenario; it’s a daily practice where an AI agent monitors an engineer’s screen, building features and even opening pull requests within hours of a discussion on Slack.

This groundbreaking initiative is powered by Block’s internally developed and open-sourced AI agent, known as “Goose.” Lenny Rachitsky recently sat down with Dhanji R. Prasanna, CTO at Block, to delve into the intricacies of this transformation and understand how Block is achieving such remarkable results.

Unprecedented Productivity Gains

The impact of Goose on productivity is staggering. Engineers utilizing the agent report saving an average of 8-10 hours per week. Extrapolating this across the entire company, including non-technical departments like support, legal, and risk, Block is realizing a monumental 20-25% reduction in manual hours. This translates to hundreds of thousands of work hours saved weekly, a testament to the power of intelligent automation.

Surprising Impact on Non-Technical Teams

Perhaps the most unexpected outcome of Goose’s implementation is the significant productivity boost experienced by non-technical teams. Dhanji highlighted the enterprise risk management team’s success in building an entire self-service system, a task that previously spanned weeks, now accomplished in mere hours. This bypasses traditional bottlenecks, such as waiting for Q2 roadmaps or relying on internal application development teams.

The Journey to Goose: From Manifesto to Open Source

Dhanji shared the comprehensive journey behind Goose’s creation and deployment. This included:

• Securing buy-in from leadership, starting with an “AI manifesto” presented to Jack Dorsey.
• Reorganizing the company structure from a General Manager (GM) model to a functional one to better support AI initiatives.
• The strategic decision to release Goose as an open-source project, fostering broader innovation and collaboration.

The full conversation with Dhanji R. Prasanna offers deep insights into Block’s AI strategy, the technical architecture of Goose, and the organizational shifts required to embrace such transformative technology. It serves as a valuable case study for any organization looking to harness the power of AI for operational efficiency and innovation.

The complete discussion is available on various platforms:

• YouTube: https://lnkd.in/gWGrWbvt
• Spotify: https://lnkd.in/gw_tVQPV
• Apple: https://lnkd.in/g5wDxe2e