Peter Yang• Apr 22, 2026• 26:12Interview

How to Build for AI Agents and a Claude Code Second Brain in 25 Min | Ryan Wiggins

Ryan Wiggins•Product Lead, Mercury

Executive Summary

Mercury has launched a Multi-modal Conversational Platform (MCP) in the Claude app store, enabling users to query their financial data using natural language to find insights like cost savings and tax breaks.
The primary interface for banking is shifting from web and mobile apps to APIs connected to conversational interfaces, representing a major platform evolution for financial services.
Internally, Mercury is accelerating product development by using AI for disposable front-end prototypes and has built an AI data analyst agent that answers 80-90% of common business queries.
The AI ecosystem is evolving with MCPs acting as a user-friendly 'app store' layer for platforms like Claude, while more technical users will be served by upcoming tools like CLIs.

12 quotes

Concerns Raised

The initial choice of an AI model provider can lead to significant and potentially restrictive platform lock-in for startups.
MCPs can be inefficient, consuming excessive context window space for power users.
Incumbent financial institutions make it difficult for users to export their own data, stifling innovation.

Opportunities Identified

Using conversational AI to analyze financial data can uncover significant, previously unknown savings and tax advantages.
APIs and MCPs represent a new, powerful distribution channel for financial products, meeting users in their existing workflows.
Building internal AI agents for tasks like data analysis can dramatically increase operational efficiency and unblock teams.
Developing personalized AI 'second brain' systems can provide a significant competitive advantage for executive decision-making.

Key Themes

The Future of Banking Interfaces

The discussion posits that the primary user interface for banking is evolving beyond websites and mobile apps. The new paradigm is centered on APIs that connect financial services to conversational AI platforms, making banking more integrated and accessible.

This signals a fundamental shift in how financial products will be built, distributed, and consumed, requiring banks and fintechs to adopt an API-first strategy to remain relevant.

AI as a Productivity Multiplier

Mercury is leveraging AI to significantly boost internal efficiency. This is demonstrated through the creation of an automated AI data analyst that democratizes data access and the use of disposable front-end prototypes which has replaced traditional spec documents, accelerating the product development lifecycle.

This provides a practical blueprint for how technology companies can integrate AI into core workflows to unblock teams, speed up innovation, and make data-driven decisions more accessible.

The Emerging AI App Ecosystem

The conversation frames MCPs as the 'third-party apps' of the AI era, providing a crucial distribution layer for services on consumer AI platforms like Claude. This is contrasted with more technical tools like CLIs, which cater to power users, illustrating a multi-layered approach to building on AI platforms.

Understanding this dynamic is critical for developers and businesses creating AI-native products, as it informs distribution strategy and how to target different user segments within the new ecosystem.

Personalized AI Augmentation

The guest details his personal 'second brain,' a custom AI system built on a 5-million-word knowledge base of his work history. This system provides contextualized insights for meetings and decisions, demonstrating a powerful application of AI for individual productivity and effectiveness.

This highlights the potential for executives and knowledge workers to build personalized AI assistants that augment their memory and decision-making, representing a new frontier in personal productivity tools.

Platform Lock-in in the AI Era

A key observation is that a startup's initial choice of a foundational AI model provider (e.g., OpenAI vs. Anthropic) can have significant downstream consequences. This initial decision influences subsequent tool adoption and creates ecosystem lock-in, making it a critical strategic choice.

This is a crucial strategic consideration for any company building on top of large language models, as the long-term costs of switching and the benefits of a cohesive ecosystem must be weighed early on.

Get started free

Topics

AI in Fintech Conversational Banking API Strategy Multi-modal Conversational Platform (MCP)Product Development AI Productivity Second Brain Knowledge Management Anthropic Claude OpenAI Startup Banking Data Analysis AI Agents Command Line Interface (CLI)Platform Lock-in

Processed Apr 22, 2026 yt-dlp + mlx-whisper + Gemini