RCM Workflow Specialist
Joyful Health, New York, NY, United States
About Joyful Health
Building the financial operating system for healthcare — and bringing the joy back to healthcare by fixing the financial chaos behind it. The healthcare payment system is a complex and inefficient maze. Healthcare practices leave $125 billion in revenue uncollected each year, lost in the chaos of fragmented financial data, manual workflows, and opaque payer systems. This financial uncertainty leaves practices struggling to stay afloat while valuable revenue slips through the cracks.
Joyful Health is building an AI‑powered financial operating system for healthcare practices. Our mission is to bring the joy back to running a private practice by simplifying financial operations so providers can focus on patient care. We spent 10 months working as fractional CFOs for a dozen practices, doing this work side by side with providers as we developed our product.
We just closed a funding round led by world‑class investors and angels including the founders of MongoDB & KAYAK.
If you’re excited about combining real‑world RCM expertise with product innovation, we’d love to meet you.
The Role
The RCM Workflow Specialist sits at the intersection of revenue cycle operations and product intelligence. This is not a traditional billing role.
You are responsible for two things that most billing professionals have never been asked to do together: 1) work real claims at an expert level and 2) translate that expertise into the structured data and direct feedback that powers Joyful’s AI product. Every action you take — every agent decision you review, every denial you analyze, every correction you document — directly improves how our system understands and solves revenue cycle problems.
You are both:
An expert‑level RCM practitioner
A collaborator who shapes how our AI learns
This role is part of the RCM Center of Excellence (CoE) and partners directly with the Engineering and Product teams.
There are two distinct tracks of work in this role:
Track 1: Agent QA & Data Labeling
Structured, recurring. You review AI agent decisions on claims, label them correct or incorrect with rationale, and provide natural language corrections that feed directly into model training. The quality of your labels is what makes the model smarter. This requires genuine RCM expertise — many decisions are judgment calls, not rule lookups.
Track 2: Product UX & Collaboration
Collaborative, fluid. You work directly alongside Engineering to pressure‑test features against real‑world RCM workflows. You tell us what’s hard to use, what doesn’t reflect how billing actually works, and what’s missing. You push back — this is an active, opinionated role, not a passive observer.
What You’ll Do
Execute High‑Quality Claims Workflows
Work claims across denial, A/R, and follow‑up workflows with a focus on accuracy and decision quality — not volume
Perform investigation, correction, and resolution of claims
Interact with payer systems, portals, and call centers as needed
Get creative when standard paths don’t work — you find a way to get the claim resolved. There is no such thing as “we can’t do it.”
Review and Label Agent Decisions
Review a statistical sample of AI agent‑closed encounters each week and assess whether the agent’s action was correct, with a written rationale for your decision
Provide natural language corrections through the product interface (e.g., “We should have checked the payer portal before closing — there was a timely filing issue that needed a retro authorization”)
These corrections feed directly into model training — your expertise is what makes the system smarter
Flag patterns where the agent consistently struggles or makes avoidable errors
Make confident calls on ambiguous scenarios — many of these are judgment calls, and sitting on the fence is not an option
Structure Data Through Labeling
Translate claim activity into standardized, structured workflow outputs
Accurately label denial categories and CARC/RARC codes
Root cause reasoning
Recovery actions taken
Outcomes (paid, denied, written off, appealed, etc.)
Ensure every claim worked and every agent review produces clean, structured data
Partner Directly with Product & Engineering
Work alongside engineers in a fast‑moving, collaborative environment — engineers communicate directly and get to the point quickly; you need to be comfortable in that environment and able to hold your own on RCM expertise
Tell us when something is wrong — if a proposed workflow doesn’t reflect how billing actually works, say so clearly and explain why
Help validate whether features and agent behaviors reflect real‑world RCM operations before they are released
Act as the voice of the RCM practitioner in product development — you are shaping how the system thinks, not just reviewing what it does
Leverage AI tools actively in your work — we expect everyone on this team to be pushing the boundary of what’s possible with the tooling available to them
Identify Patterns & Surface Insights
Recognize trends across claims, payers, and denial types
Flag inconsistencies, contradictions, or unclear outcomes
Surface edge cases and breakdowns in workflows
Contribute to improving categorization logic, definitions, and SOP quality
How You’ll Be Measured
This role is measured on quality, not quantity. Throughput is not a primary metric — the goal is expert‑level accuracy and meaningful insight generation.
Accuracy of data labels and agent QA decisions
Quality and confidence of rationale provided for ambiguous cases
Quality and specificity of product feedback — not just “this is broken” but “here’s why and here’s what should happen instead”
Consistency of workflow execution
Signal contribution to product and model improvements — measurable instances where your feedback changed something for the better
Independence and confidence on judgment calls — a key marker of success is that you don’t need to escalation ambiguous cases — you make a call and explain it
What Success Looks Like
In 30 Days
Deep familiarity with Joyful workflows, the product interface, and labeling structure
Producing accurate, well‑reasoned claim documentation and agent QA labels
Comfortable navigating payer systems and making independent decisions on clear‑cut cases
Establishing a working relationship with the Engineering team — communication style calibrated, expectations clear
In 90 Days
Independently working claims and QA‑ing agent decisions across multiple denial categories
Consistently producing high‑quality labeled data with clear, confident rationale
#J-18808-Ljbffr
Building the financial operating system for healthcare — and bringing the joy back to healthcare by fixing the financial chaos behind it. The healthcare payment system is a complex and inefficient maze. Healthcare practices leave $125 billion in revenue uncollected each year, lost in the chaos of fragmented financial data, manual workflows, and opaque payer systems. This financial uncertainty leaves practices struggling to stay afloat while valuable revenue slips through the cracks.
Joyful Health is building an AI‑powered financial operating system for healthcare practices. Our mission is to bring the joy back to running a private practice by simplifying financial operations so providers can focus on patient care. We spent 10 months working as fractional CFOs for a dozen practices, doing this work side by side with providers as we developed our product.
We just closed a funding round led by world‑class investors and angels including the founders of MongoDB & KAYAK.
If you’re excited about combining real‑world RCM expertise with product innovation, we’d love to meet you.
The Role
The RCM Workflow Specialist sits at the intersection of revenue cycle operations and product intelligence. This is not a traditional billing role.
You are responsible for two things that most billing professionals have never been asked to do together: 1) work real claims at an expert level and 2) translate that expertise into the structured data and direct feedback that powers Joyful’s AI product. Every action you take — every agent decision you review, every denial you analyze, every correction you document — directly improves how our system understands and solves revenue cycle problems.
You are both:
An expert‑level RCM practitioner
A collaborator who shapes how our AI learns
This role is part of the RCM Center of Excellence (CoE) and partners directly with the Engineering and Product teams.
There are two distinct tracks of work in this role:
Track 1: Agent QA & Data Labeling
Structured, recurring. You review AI agent decisions on claims, label them correct or incorrect with rationale, and provide natural language corrections that feed directly into model training. The quality of your labels is what makes the model smarter. This requires genuine RCM expertise — many decisions are judgment calls, not rule lookups.
Track 2: Product UX & Collaboration
Collaborative, fluid. You work directly alongside Engineering to pressure‑test features against real‑world RCM workflows. You tell us what’s hard to use, what doesn’t reflect how billing actually works, and what’s missing. You push back — this is an active, opinionated role, not a passive observer.
What You’ll Do
Execute High‑Quality Claims Workflows
Work claims across denial, A/R, and follow‑up workflows with a focus on accuracy and decision quality — not volume
Perform investigation, correction, and resolution of claims
Interact with payer systems, portals, and call centers as needed
Get creative when standard paths don’t work — you find a way to get the claim resolved. There is no such thing as “we can’t do it.”
Review and Label Agent Decisions
Review a statistical sample of AI agent‑closed encounters each week and assess whether the agent’s action was correct, with a written rationale for your decision
Provide natural language corrections through the product interface (e.g., “We should have checked the payer portal before closing — there was a timely filing issue that needed a retro authorization”)
These corrections feed directly into model training — your expertise is what makes the system smarter
Flag patterns where the agent consistently struggles or makes avoidable errors
Make confident calls on ambiguous scenarios — many of these are judgment calls, and sitting on the fence is not an option
Structure Data Through Labeling
Translate claim activity into standardized, structured workflow outputs
Accurately label denial categories and CARC/RARC codes
Root cause reasoning
Recovery actions taken
Outcomes (paid, denied, written off, appealed, etc.)
Ensure every claim worked and every agent review produces clean, structured data
Partner Directly with Product & Engineering
Work alongside engineers in a fast‑moving, collaborative environment — engineers communicate directly and get to the point quickly; you need to be comfortable in that environment and able to hold your own on RCM expertise
Tell us when something is wrong — if a proposed workflow doesn’t reflect how billing actually works, say so clearly and explain why
Help validate whether features and agent behaviors reflect real‑world RCM operations before they are released
Act as the voice of the RCM practitioner in product development — you are shaping how the system thinks, not just reviewing what it does
Leverage AI tools actively in your work — we expect everyone on this team to be pushing the boundary of what’s possible with the tooling available to them
Identify Patterns & Surface Insights
Recognize trends across claims, payers, and denial types
Flag inconsistencies, contradictions, or unclear outcomes
Surface edge cases and breakdowns in workflows
Contribute to improving categorization logic, definitions, and SOP quality
How You’ll Be Measured
This role is measured on quality, not quantity. Throughput is not a primary metric — the goal is expert‑level accuracy and meaningful insight generation.
Accuracy of data labels and agent QA decisions
Quality and confidence of rationale provided for ambiguous cases
Quality and specificity of product feedback — not just “this is broken” but “here’s why and here’s what should happen instead”
Consistency of workflow execution
Signal contribution to product and model improvements — measurable instances where your feedback changed something for the better
Independence and confidence on judgment calls — a key marker of success is that you don’t need to escalation ambiguous cases — you make a call and explain it
What Success Looks Like
In 30 Days
Deep familiarity with Joyful workflows, the product interface, and labeling structure
Producing accurate, well‑reasoned claim documentation and agent QA labels
Comfortable navigating payer systems and making independent decisions on clear‑cut cases
Establishing a working relationship with the Engineering team — communication style calibrated, expectations clear
In 90 Days
Independently working claims and QA‑ing agent decisions across multiple denial categories
Consistently producing high‑quality labeled data with clear, confident rationale
#J-18808-Ljbffr