How Block Rapidly Hired 12,000 Employees with AI

How Block Got 12,000 Employees Using AI Agents in Two Months

In an era where many companies are still grappling with the integration of artificial intelligence (AI) into their development processes, fintech company Block has made headlines by successfully deploying AI agents to its entire workforce of 12,000 employees in just eight weeks. This remarkable achievement was shared by Angie Jones, Block’s vice president of engineering, during the All Things Open 2025 conference held in Raleigh, N.C.

It Started Small

The story of Block's rapid deployment began with a single engineer’s frustration. Bradley, a principal machine learning (ML) engineer at Block, was dissatisfied with existing AI tools that only generated code snippets. He sought a solution that could automate complex development tasks, rather than simply suggesting the next line of code.

When OpenAI introduced function calling, Bradley and a small team began experimenting with automating parts of their development workflow. Their initial experiments proved successful, but they soon encountered a scaling problem. The lack of standardized procedures made it difficult to integrate various APIs, resulting in a maintenance nightmare. Each new tool required custom code, leading to inefficiencies.

The Role of the Model Context Protocol

The breakthrough came when Anthropic approached Block with the Model Context Protocol (MCP), an open standard designed for connecting AI agents to tools and data sources. Recognizing the potential of this protocol, Block became a launch partner and revised their internal AI agent, dubbed “Goose,” to operate within the MCP framework.

The Accidental Discovery

While developing Goose primarily for engineers, the team had an epiphany: they could adapt it for use by all 12,000 employees. Angie Jones recounted the moment of realization, “We started thinking, what if we could make this work for all 12,000 employees?” However, this ambition was not without challenges.

Challenges in Implementation

Initially, attempts to broaden the use of Goose were met with significant obstacles. Most of the workforce had little to no experience with command-line interfaces or managing API keys. Early users struggled to install the software, while others who succeeded faced daunting authentication flows. The user experience was primarily designed for developers, which posed a barrier for non-technical staff.

Making It Work for Real People

Recognizing these challenges, the team shifted their focus to enhance user experience. They began to simplify the installation process and streamline authentication flows to cater to non-technical employees. This involved creating user-friendly interfaces and providing extensive training and support.

• User Training: Block organized workshops and training sessions to familiarize employees with Goose and its capabilities.
• Feedback Mechanism: They established a feedback loop, allowing employees to report issues and suggest improvements, making the tool more user-centric.
• Documentation: Comprehensive documentation was created to assist employees in navigating the software, covering everything from installation to troubleshooting.

Practical Implications and Use Cases

The deployment of AI agents like Goose has had profound implications for Block. By enabling all employees to leverage AI, the company has increased productivity and efficiency across various departments. Here are some practical examples of how Goose is being utilized:

• Customer Support: Customer service representatives can use Goose to automate responses to frequently asked questions, reducing response times and improving customer satisfaction.
• Data Analysis: Employees in finance can utilize Goose to analyze large datasets quickly, generating insights that would take much longer using traditional methods.
• Marketing Automation: Marketing teams can automate campaign management tasks, from scheduling posts to analyzing engagement metrics, allowing them to focus on strategy rather than execution.

Conclusion

Block's rapid deployment of AI agents to its workforce demonstrates the transformative power of AI when implemented thoughtfully. By leveraging the Model Context Protocol and focusing on user experience, Block successfully turned a complex tool into an accessible resource for all employees. This case serves as a valuable lesson for other organizations looking to harness AI technology: understanding user needs and providing adequate support is crucial for successful implementation.

As AI continues to evolve, companies that prioritize inclusivity in their technology adoption will likely find themselves at the forefront of innovation. For more details on the Model Context Protocol and its applications, you can visit the official MCP page.