Financial Planners Build and Review Their Own AI Skills

Matt Ragone, a financial planner at Range, describes how his team builds AI-assisted workflows without a coding background.

Start with the work

Range had already documented how its planners make insurance, investment and retirement recommendations.
Its connected client data gives the agent material to retrieve; the planner must still check that it used the right information.
Choose a frequent, relatively simple workflow whose output an expert can judge, such as a cash-flow or insurance review.

Teach: supply real examples and explain decisions, including edge cases and past corrections.
Pack: turn the process into a reusable skill and have it ask for missing context, such as the tax year or review focus.
Set limits: Ragone says a certified financial planner reviews each skill’s output before it goes to a client.
Test: try five to ten real cases, including edge cases, and involve colleagues in stress testing.
Ship with an owner and reviewer: keep updating the skill as tax rules and business practices change.

Ragone reports about 24 skills across the company, including 17 in financial planning, and describes an internal Slack agent still under development.
Subject specialists help build and review the skills in their own domains; an AI council shares practices across teams.
His ambition to serve 100 people a day per planner is a goal, not a measured result. He offers no controlled productivity comparison.

“It took a lot more than just a good prompt.”