Block deployed the Builderbot agent framework across its codebase to automate complex cross-service software development tasks.
Block engineers executed a complete operational overhaul of their internal development pipelines over the past 24 months. The engineering department focused intensive investment on integrating AI into the standard daily routines of every technical staff member. This corporate initiative resulted in total adoption, with all engineers actively using these internal tools in their regular programming duties.
Standard single-repository coding assistants quickly demonstrated limitations during this internal rollout. These basic applications failed to process the true mechanical complexity of a mature corporate software environment containing hundreds of interconnected active services. The internal corporate networks store hundreds of millions of lines of proprietary code. Block constructed the Builderbot application explicitly to solve these mass implementation blockages and provide genuine cross-system execution capabilities.
Builderbot operates as a highly active central orchestration layer. The application coordinates multiple distinct software agents to manage complex technical assignments across the infrastructure.
Engineers initiate the entire deployment sequence directly within standard Slack communication channels. A developer tags the @builderbot account and provides a concise text description of the required technical operation. The agent instantly begins processing the request within the exact same chat thread. The requested parameters vary wildly in structural complexity, ranging from basic software bug resolutions to massive architectural migrations running across multiple internal databases.
Collaborative platform engineering takes place entirely within the primary communication platform. Several team members can monitor the active operational thread simultaneously, observing the designated agent as it executes its internal research and planning phases. The human operators provide continuous directional steering to correct the machine’s execution trajectory. The conversation thread effectively replaces the IDE interface. Developers avoid the cognitive penalty of switching contexts between external coding programs and internal team chat platforms.
Builderbot’s agentic operations
Scale of execution defines the core difference between basic code generation systems and the Block deployment model. Builderbot actively maps the complete structural context of the entire corporate codebase—cataloguing every active network service, documenting every internal API endpoint, and internalising strict company engineering conventions.
Builderbot possesses the required digital permissions and contextual understanding to modify any software repository managed by the company. A software engineer attached to the company’s Cash App product division can trigger direct operational modifications to a separate Square backend service. The developer requires zero prior experience with the Square subsystem because the centralised orchestration layer supplies the correct architectural context automatically.
The deployment also integrates deeply into standard issue tracking and project management pipelines. Builderbot interfaces directly with project management databases, actively retrieving assigned tickets from both Linear and Jira interfaces. The autonomous system claims the ticket, provisions the initial code branch, and generates the necessary raw source code. The system then opens formal pull requests for the completed modifications on the network.
The workflow remains entirely autonomous as the system continuously monitors the automated continuous integration testing suite. Builderbot processes any automated testing failures or direct human technical feedback and iterates upon the raw code until it successfully meets production deployment standards. Human engineers insert themselves into the process strictly to provide high-level analytical judgments rather than performing manual keyboard typing.
Data security and automated pull requests
Block engineered Builderbot to operate exclusively within the controlled domain of source code repositories and structural system configurations. The network architecture strictly prevents the agent from reading, processing, or transmitting raw customer data. Most importantly, it contains zero technical access to live payment information or personally identifiable information stored within production servers. This isolation protects internal compliance structures while allowing the application complete freedom to manipulate the underlying software architecture securely.
Volume metrics from the active live deployment validate the strategy. The centralised integration layer successfully executes over 200,000 separate operational commands every single day. The automated programming system merges approximately 1,500 distinct pull requests into the production codebase on a weekly basis.
These autonomous code contributions represent roughly fifteen percent of all structural modifications occurring across the entire company network. Product development cycles compressed drastically following the system launch, collapsing initial project timelines from several expected months down to periods of days.
Brad Axen, Head of AI Capabilities at Block, said: “The best way to think about Builderbot is as the missing layer between AI coding tools and how engineering actually works at scale. It handles the orchestration, the context, the environment, so our engineers can focus on the problems worth solving.
“On the Square side, we took a list of features sellers had been waiting on for months and our engineers shipped them in days. Builderbot handled the scaffolding and the repetitive work, and our engineers made the decisions that shaped the product. It means an idea can go from backlog to live in front of millions of customers in days instead of months.”
Goose framework and Model Context Protocol
The proprietary orchestration layer sits atop public technological foundations. Block engineered the internal deployment using its open-source goose agentic framework. The company originally developed the goose framework internally before contributing the underlying code base to the Agentic AI Foundation.
The technical obstacles Block engineers encountered while attempting to implement goose across their internal networks directly prompted external industry partnerships. The internal platform integration struggles inspired a direct corporate collaboration with Anthropic.
Block and Anthropic successfully co-developed the Model Context Protocol to solve the internal data connectivity problem. The resulting technology serves as the established industry standard mechanism for linking autonomous operational agents to internal development tools and raw data sources.
Block engineering leadership stated the primary motivation for building the internal orchestration layer was raw operational necessity. The internal engineering department struggled constantly with coordinating autonomous applications across massive internal repositories while attempting to maintain high code quality at extreme development velocities.
The organisation views the technical transition from basic assisted code generation toward native agentic engineering as a central priority in the global software sector. Block published the internal details of the Builderbot system to contribute openly to this wider technical progression.
See also: Xiaomi MiMo Code executes 200-step agentic developer workflows
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information.
Developer is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.