Healthcare operations
Medical revenue cycle management: the steps, and where AI may help
A plain guide for practice managers, billing leads and health system finance teams. It says where AI could help and what stays with people.
Medical revenue cycle management is the work of getting a provider paid for care, from the first patient contact to the last collected balance. The Healthcare Financial Management Association's revenue cycle page lists patient access, billing, coding, claims management, collections, denials prevention and revenue cycle optimization. Each step passes information to the next, so a mistake early on, such as a wrong insurance detail, shows up later as a denied claim. AI may help with drafting and sorting work, but people should still decide on coding, appeals and anything that changes a bill. A good starting point is one step with a clear problem.
plain answer
What is medical revenue cycle management?
It is the set of tasks a provider uses to get paid: registering the patient and checking insurance, recording the care, coding it, sending the claim, handling denials and collecting what is owed.
HFMA's revenue cycle page, read on 11 November 2026, lists patient access, billing, coding, claims management, collections, denials prevention and revenue cycle optimization. Our steps below follow a common order and are not HFMA's list.
framework
The main steps of the revenue cycle
Steps and names vary between organizations. This is a common order.
1. Patient access
Scheduling, registration and checking insurance details before the visit.
2. Coding
Turning the care recorded in the chart into the codes a payer expects.
3. Billing and claims
Building the claim and sending it to the payer.
4. Denials
Working out why a payer refused or reduced a claim, and fixing or appealing it.
5. Collections
Following up on balances owed by payers and patients.
6. Review
Looking at where claims stall so the cause is fixed upstream.
how to
Where AI may help and where people decide
These are our own examples of sensible splits, not outcomes we have measured.
| Step | What AI may help with | What people decide |
|---|---|---|
| Patient access | Flagging missing or mismatched insurance fields for review. | Whether the details are right. |
| Coding | Suggesting codes and flagging likely errors before submission. | The final code. |
| Denials | Sorting denials by stated reason and drafting appeal letters. | Whether to appeal and what is sent. |
| Collections | Drafting reminders and sorting accounts for follow up. | Who is contacted, and how. |
| Review | Summarizing where claims stall. | What to change. |
differentiation
Start with one step, not the whole cycle
Because each step feeds the next, it is tempting to redesign all of it. A smaller start is to pick the one step where work piles up, write down what a person checks now and test help there. Records that include patient details need an agreed access rule before any AI sees them. Our AI vendor evaluation checklist lists questions to ask any vendor first.
For rules on what data AI may use, see AI data governance. For automation across billing and claims workflows more widely, see healthcare workflow automation and AI governance.
kriv role
Where Kriv helps
Kriv AI is an AI consultancy for regulated organizations, with HIPAA-aware delivery. We help health care teams decide where AI fits, what is logged and who reviews it. Our revenue cycle accelerator page describes a design on Google Cloud for denial classification, coding support and appeal drafting. That page states that the architecture and data schema are designed, but the model training pipeline and analytics dashboard are not yet built, and that it uses synthetic data only.
We do not run billing for you, we do not submit claims and we do not make coding decisions. Work is scoped and priced after a short discovery call.
limits
What this page does not cover
This page explains the process in general terms. It is not billing, coding or legal advice, and payer rules differ. We cite no denial rates or savings figures, because we did not find a source we could read and quote on 11 November 2026.
Straight answers
Frequently asked questions about Medical revenue cycle management: the steps, and where AI may help
What is medical revenue cycle management?
It is the work of getting a provider paid for care, from patient access and coding through claims, denials and collections.
What are the steps of the revenue cycle?
A common order is patient access, coding, billing and claims, denials, collections and review. HFMA's revenue cycle page lists patient access, billing, coding, claims management, collections, denials prevention and revenue cycle optimization.
Can AI do medical billing?
AI may help draft, sort and flag work, but people should decide on coding, appeals and anything that changes a bill.
Has Kriv AI built revenue cycle AI?
Our accelerator page describes a design on Google Cloud. It says the architecture and data schema are designed and the model pipeline and dashboard are not yet built, using synthetic data only.
Where should a practice start?
Pick the one step where work piles up, write down what a person checks now and test help there before widening.
Talk to the team that would do the work
Bring your requirements to a working session with the person who'll actually deliver.
Book a Discovery Call