Teamday
← Back

Range: Financial Planners With No Code Built 17 AI Skills

AI Engineer
agentsenterprisefuture-of-workknowledge-workbusiness

How Range's Financial Planners Built Their Own AI Agents

Matt Ragone is a financial planner at Range, an all-in-one wealth management firm and registered investment adviser, with about seven years in the industry. He also leads AI adoption for Range's planning team through the company's AI Council, a roundtable of the most advanced AI users from every department, set up by co-founder David Cusatis so teams can share skills instead of rebuilding them. His AI Engineer talk is a build guide from people who, in his words, are new to this: "And none of us have any engineering, coding background."

The bottleneck was typing speed: Early on the team used AI as a chatbot for redrafting emails, checking tax code changes and parsing long documents. It was faster, but planners still did all the work by hand, and a 45-minute client call meant no work for anyone else. Claude Cowork changed that in early 2026. While Ragone is on a call with one client, the agent drafts a cash flow analysis, an insurance review or a tax review for another. His job shifts to reviewing and adjusting the agent's work until it is bulletproof before it goes out.

Standardization came first: Two things had to be true. The agent had to work where client data lives, since Range members connect their accounts, insurance, investments and mortgages to the platform. And it had to know how Range gives advice, which is more conservative than a generic agent would be. The 20-person planning team had already written down how it delivers insurance, investment and retirement recommendations, so the agent could simply be told to reference those documents before any workflow.

Five steps, no code: Range's method is pick, teach, pack, test and ship. Pick a workflow repeated daily or weekly, not the flashiest one, because you can judge the output right away. Teach with real cases, now two to five rather than one, including edge cases where a planner corrected the agent, and narrate why each input matters, for example where to find a group life insurance deduction on a pay stub. Pack it into a skill that asks clarifying questions first (which tax year, which state, homeowner's or disability focus) and states its limits. Test on five to ten real cases with colleagues stress-testing it, then ask the agent how it would improve itself. Ship it with a named owner and a separate reviewer. "So that's why the people doing the work should be the ones building it", Ragone says, with insurance specialists owning the insurance skill and the CPA team shaping the tax skills.

Regulated output stays human-reviewed: Because the output is financial advice to real clients, "So every single skill is not allowed to send anything to a member, okay?" A certified financial planner reviews everything before it reaches a client, and Ragone expects that to remain the case.

The results after six to eight months: "And we have about 24 skills across the company, 17 just in financial planning." Planners, marketers and the member success team build on each other's skills, and Range is now building Ranger, an internal agent in Slack that helps everyone at the company with tasks.

5 Takeaways From Range on Non-Engineers Building AI Agents

  • Standardize the work before automating it - Written recommendation standards became the agent's reference material on day one
  • Start with boring, frequent workflows - Cash flow analyses and insurance reviews beat complex showcase projects because quality is easy to judge
  • Teach with corrected real cases - Edge cases where a planner fixed the agent's work make the next run better
  • Every skill needs an owner and a reviewer - Subject-matter experts own the content; tax law and products keep changing
  • Hard limits for regulated output - No skill sends anything to a client without CFP review

What Range Shows About the Future of Financial Advice Work

The bottleneck has moved from how fast a planner can type to who keeps every skill tested and current. Ragone's prediction: "We're almost going to start having planners, I would imagine, almost be AI agent managers, right?" His stated goal is moving each planner from advising three or four people a day to a hundred, with the experts who do the work writing the skills that do it.