# How Financial Planners With Zero Coding Experience Built Their Own Agents - Matt Ragone, Range

## Executive summary

Matt Ragone, a financial planner at Range, details a non-technical, five-step methodology for building and deploying AI agents and skills. The process allows domain experts, rather than engineers, to automate complex, repetitive tasks (like cash flow or tax reviews) using tools like Claude Co-work. The core principles emphasize standardizing workflows, centralizing client data, and establishing mandatory human review points (CFP/RIA compliance) to ensure accuracy and compliance.

## Key takeaways

- The Shift from Chatbots to Agents: Initial AI use was limited to chatbots for drafting emails or fact-checking. The major advancement came with agents (like Claude Co-work) that could perform complex tasks (e.g., drafting cash flow analyses, tax reviews) concurrently while the planner was servicing other clients, moving beyond manual, single-task limitations. (3:37)
- The Five-Step Skill Building Process: The methodology for building an agent skill is: 1) Pick a workflow you repeat constantly (start simple); 2) Show it real cases (including edge cases); 3) Package it as a skill (requiring defined questions and limits); 4) Test it on 5-10 real cases; and 5) Ship it with a designated owner and a reviewer. (6:27)
- The Importance of Domain Expertise: The speaker emphasizes that the people who perform the work should be the ones building the skill. This ensures the agent understands the nuances and edge cases of the domain (e.g., an insurance specialist iterating on the insurance skill). (14:28)

## Technical details

- Data Architecture and Integration: For an agent to be effective, it must access centralized client data (e.g., insurance documents, investment accounts) from the firm's platform, rather than relying on data copied and pasted into a chat window. The agent must know how to correctly pull and interpret this structured data. (4:21)
- Agent Compliance and Review: Because the firm operates as a regulated RIA, all skills must adhere to strict compliance standards. Critically, every skill is prohibited from sending advice directly to a member; it must be reviewed by a Certified Financial Planner (CFP) before deployment. (3:03, 10:50)
- Skill Packaging Requirements: When packaging a skill, the agent must be prompted to ask clarifying questions (e.g., 'Is this a 2025 tax review?') and must be explicitly given defined limits and constraints to respect the firm's conservative advice standards. (11:52)

## Practical implications

- The process demonstrates that complex AI automation can be democratized, requiring process standardization and domain knowledge rather than deep coding expertise.
- Build teams should focus on creating standardized, repeatable workflows and defining clear data access layers to enable non-technical subject matter experts to build and iterate on automation skills.
- Implementing a mandatory 'Owner' and 'Reviewer' structure for every deployed skill is critical for maintaining compliance and quality control in regulated industries.

## Topics

AI Agents, Workflow Automation, Financial Planning, Process Standardization, RIA Compliance, No-Code AI Development, Range, Claude Co-work, AIE CODE summit

Source: https://www.youtube.com/watch?v=EaOit1mmR9g
