Podcast Summary: How I AI

Episode: How Block’s Custom AI Agent Supercharges Every Team, from Sales to Data to Engineering
Guests: Jackie Brosamer & Brad Axen
Release Date: July 28, 2025
Host: Claire Vo

Introduction

In the July 28, 2025 episode of How I AI, host Claire Vo engages in an insightful discussion with Jackie Brosamer and Brad Axen from Block. The episode delves into how Block has seamlessly integrated AI across various departments, leveraging their open-source AI agent, Goose, to enhance productivity and foster organizational transformation.

Embracing AI Across the Organization

[03:03] Claire Vo: Claire opens the conversation by commending Block for its comprehensive and swift adoption of AI, contrasting it with the common narrative of AI-native startups emerging rapidly. She remarks, "It seems everywhere and for everything," highlighting Block's holistic approach to AI integration.

[03:03 - 04:49] Jackie Brosamer: Jackie attributes Block's successful AI adoption to a combination of bottom-up and top-down strategies. She states, "We saw a lot of concentrated pockets once ChatGPT came out that were really finding a lot of value out of these tools." This dual approach ensured that both frontline teams like sales and technical teams like engineering embraced AI, driving consistent and widespread usage across the company.

[04:49] Claire Vo: Claire emphasizes the cultural and operational transformations driving Block’s AI success, noting, "Not only are you all leading the way in terms of how organizations... are trying to drive and help others adopt it externally."

Introduction to Goose: Block’s Open-Source AI Agent

[05:18] Brad Axen: Brad introduces Goose, describing it as an AI agent powered by a Large Language Model (LLM) and a suite of tools that enable problem-solving across diverse tasks. He explains, "You tell it what you needed to do by connecting it to different capabilities and it can just kind of solve any problem."

[06:40] Jackie Brosamer: Jackie underscores Block's commitment to open source, explaining, "Open source is, I think, a really big emphasis just as a value for our company." By open-sourcing Goose, Block fosters an ecosystem where developers worldwide can contribute, extending Goose's functionalities and integrating various models to enhance its versatility.

Demonstration: Goose in Action

Vibe Data Analysis

[08:13] Jackie Brosamer: Jackie presents a practical use case by analyzing a CSV containing July data from her small farm's stand. She demonstrates Goose’s ability to employ Python’s Pandas library to extract insights such as top revenue-generating items and busiest sales days.

Notable Quote:

"What a lot of my work and what my team is trying to do is not just make it faster for someone like me... but to allow our finance team, our sales team, anyone in the company to be able to dig in and self-serve a lot of this data rather than having to ask an expert and wait for that to come back." — Jackie Brosamer [00:30]

[09:55] Claire Vo: Claire marvels at Goose's capability to handle complex data tasks effortlessly, likening it to bypassing manual spreadsheet manipulations by saying, "Is this just like general day to day sales of different products?"

Building and Extending MCPs

[17:31] Brad Axen: Brad transitions to demonstrate Goose’s coding capabilities, focusing on creating Model Context Protocols (MCPs). He explains, "Each of these is an MCP server... it's like a protocol that lets the agent talk to a third party and get things that it can do and data it can read."

[18:05 - 32:41] Live Coding Session: Brad showcases Goose's ability to build an MCP live. Starting with integrating Goose with Square’s API, he guides the audience through importing data, handling Python dependencies, and generating secure email functionalities. The session highlights Goose's proficiency in troubleshooting, as it autonomously identifies and resolves errors during the coding process.

Notable Quote:

"The debugging flow is like something that I thought was going to be what kept me employed with the AI writing all these code, but I'm like, oh, it's actually pretty good." — Brad Axen [38:44]

Organizational Impact and Insights

[15:36] Brad Axen: Brad discusses the versatility of MCPs, mentioning, "We have like a collection of different MCPs. Some of them we've taken from the open source community, some of them we've developed internally." This adaptability allows Goose to connect seamlessly with various internal systems like Google Drive and external platforms, enhancing its utility across departments.

[17:05] Claire Vo: Claire reflects on Goose’s ability to transition from simple data analysis to more complex tasks like generating payment links, illustrating the AI agent's scalability and practical application in real-world scenarios.

[17:31 - 43:35] Continuous Development and User Engagement: The conversation underscores the importance of user-friendly AI tools in democratizing data access and empowering non-technical teams. Jackie notes, "We've all been talking about like the age of self-serve data... to allow everybody to focus less on the rote parts of data analysis and spend more time on the insights and the context."

Lightning Round: Favorites and Tips

Favorite MCP Tools:

• Jackie Brosamer: "Google Drive. That is, as a manager where I live, and I'm just so happy to be able to not have to copy and paste and click and save me a ton of time." [42:35]
• Brad Axen: "Definitely shell. I just, you know, I've lost... it's just 100% filling them in for me at this point." [42:35]

Advice for AI Hesitants:

• Brad Axen: "Find the thing that you don't like doing and automate that." [43:30]
• Jackie Brosamer: "Treating it as a learning experience and realizing that this is the worst the models are ever going to be." [43:35]

Handling AI Frustrations:

• Brad Axen: Emphasizes the importance of restarting sessions when Goose misbehaves, stating, "I just throw out the whole combo and I start over." [44:23]
• Jackie Brosamer: Suggests summarizing sessions and refining instructions, saying, "Take out the bad parts and start a new session based on those instructions." [44:58]

Conclusion and Next Steps

Claire wraps up the episode by reflecting on the comprehensive capabilities demonstrated by Goose, from data analysis to live MCP development. She encourages listeners to explore Goose and integrate similar AI-driven tools within their organizations to enhance efficiency and foster innovation.

[45:35] Brad Axen: Provides resources for listeners to access Goose, directing them to Block’s GitHub repository and inviting contributions and discussions via their Discord channel.

Closing Thoughts: The episode underscores the transformative potential of AI agents like Goose in bridging technical and non-technical workflows, promoting a culture of innovation, and enabling self-service data capabilities across organizations.

Key Takeaways:

• Holistic AI Integration: Block's bottom-up and top-down approach ensures widespread AI adoption across departments.
• Open-Source Collaboration: Goose’s open-source nature fosters community-driven enhancements and diverse integrations.
• Empowering Non-Experts: Tools like Goose democratize data access, enabling various teams to perform complex analyses without deep technical expertise.
• AI-Assisted Development: Goose streamlines coding processes, automates repetitive tasks, and assists in building extensible protocols (MCPs).
• Cultural Transformation: Emphasizing operational and cultural shifts is crucial in leveraging AI for organizational success.

Resources:

• Goose GitHub Repository: Block GitHub IO Goose
• Block Open Source Discord: Accessible through the GitHub repository for real-time discussions and support.

For more episodes and insights on integrating AI into your workflow, subscribe to How I AI on Apple Podcasts, Spotify, or your preferred podcast platform.