Background
A large, integrated healthcare system serving communities across the Carolinas, explores opportunities to leverage AI and automation. Our client needed additional expertise and resources to identify high-impact initiatives, optimize existing Epic workflows, and evaluate types of AI architecture that could improve efficiency and create measurable financial value through AI and automation.
The Challenge
AI and automation are helpful tools to relieve administrative burden, improve revenue performance and help address existing staffing gaps. However, layering new AI technology onto an unoptimized Epic build could lead to automation of already existing problems within an organization. Before adding any new technology, health systems need to make sure their existing workflows, data, and Epic configuration has the ability to support it.
Our client manages a large set of processes, from patient scheduling and registration of patients to claims management and reimbursement. Many of these workflows require significant manual labor and depend on staff finding and interpreting changing payer requirements. Investing in AI and automation isn’t enough to completely transform revenue cycle operations.
And before moving forward with this project our client needed to understand which processes should be improved, whether its already existing Epic configuration could support these changes and where AI and automation could help deliver the best operation and financial impact.
That’s where HDM comes in!
The Solution
Discovery: Identifying and Prioritizing Opportunities
The engagement began with an evaluation of revenue cycle workflows, existing Epic functionality and areas of significant manual effort. HDM collaborated with revenue cycle leadership to investigate approximately 40 potential AI and automation opportunities before prioritizing the two most valuable initiatives.
Phase 1: Building the Data and Analytical Foundation
Coordination of Benefits (COB)
The first initiative focused on registration issues requiring direct patient involvement, particularly Coordination of Benefits forms.
Payer real-time eligibility responses can be inconsistent and difficult to interpret. Staff must often search payer websites, determine which form is required, and contact the patient to obtain missing information which can sometimes delay claim submission or reimbursement.
HDM analyzed eligibility data to identify active Coordination of Benefits errors, determine the payer involved, and recommend the appropriate next action. The solution was designed to:
- Identify accounts with active Coordination of Benefits errors
- Interpret payer information returned through eligibility responses
- Exclude false positives and accounts that do not require outreach
- Identify patients with active CARC 22 codes
- Detect accident-related accounts handled through other established processes
- Determine each patient’s language and communication preferences
- Identify the correct payer form or digital resource
- Prioritize issues occurring within seven days of service
This work created the data architecture, pipelines, and decision logic needed to identify which patients require outreach, what information is needed, and how each request should be delivered.
Underpayment Identification and Risk
The second initiative focused on distinguishing true underpayments from other payment variances and helping staff prioritize the accounts with the greatest financial opportunity.
HDM developed analytical capabilities that aggregate underpayment trends by payer, procedure, diagnosis, and department specialty. The solution filters out false positives, including minor rounding differences, unmapped account variances, and denial-related variances, so staff can concentrate on actionable cases.
The initial capabilities allow revenue cycle teams to:
- Identify recurring underpayment trends by payer, CPT, DRG, and specialty
- Quantify underpayment volume and financial exposure
- Calculate risk using payer, procedure, diagnosis, adjustment, and billing activity
- Detect emerging payer-specific patterns
- Exclude non-actionable variances from analysis
- Support earlier intervention for high-risk service lines
By consolidating this information, the organization can reduce the manual review of scattered, low-dollar variances while identifying patterns that may warrant escalation.
Phase 2: Activating Patient Outreach and Integrating Workflows
Automating Patient Communication
Patients with active MyChart accounts can receive digital forms that are partially prepopulated with their current insurance information and can be completed electronically. Patients who are not active in MyChart can receive a secure text message directing them to the required information. Traditional outreach methods remain available when digital delivery is unsuccessful.
The solution is also designed to:
- Deliver messages through the patient’s preferred channel
- Honor language preferences and text-message opt-outs
- Monitor unsuccessful communication attempts
- Confirm whether forms were processed through payer portals
- Add or remove Epic Coordination of Benefits flags as payer information changes
- Prevent repeated requests after a patient has completed the necessary action
- Escalate unresolved cases for additional support when appropriate
Identifying Underpayment Sources and Recommended Actions
The solution compares procedure, diagnosis, and revenue codes against contract terms, payer manuals, policies, and newsletters. This allows the revenue cycle team to identify conflicting requirements and distinguish whether a variance originated from a contract, payer policy, billing activity, or another source.
New insights are then incorporated into staff workflows to help teams:
- Identify the likely source of an underpayment
- Review relevant payer rules and supporting documentation
- Prioritize accounts with the highest probability of recovery
- Determine whether an appeal, reconsideration, managed-care escalation or legal review is most appropriate
- Analyze the sequence of billing activity leading to the variance
- Provide evidence-backed support for recurring payer escalations
- Focus staff time on claims with the strongest recovery potential
Anticipated Impact
HDM and the health system will track operational and financial performance as the capabilities move into production, since financial results are not yet available.
Expected benefits include:
- Less manual interpretation of payer eligibility responses and policies
- Faster identification and resolution of coverage issues
- Fewer unnecessary or duplicative patient communications
- Higher digital form-completion rates
- Reduced delays caused by missing Coordination of Benefits information
- More accurate identification of true underpayments
- Less staff time spent reviewing false positives and low-value variances
- Better prioritization of accounts with a high probability of recovery
- Stronger collaboration across patient access, billing, coding, and managed-care teams
- Improved visibility into payer-specific trends and reimbursement risks
Looking Ahead
Future phases will expand the solutions with additional payer, market, and regional intelligence. These capabilities could provide staff with more sophisticated guidance, support payer negotiations, and help the organization identify emerging reimbursement risks earlier.
By first strengthening its data, workflows, and Epic foundation, the health system is building a sustainable approach to AI adoption, one that supports patients and staff while creating operational and financial value.
Don’t automate your problems. Fix the foundation first, then let HDM help you find your highest-impact AI and automation opportunities. Email us at solutions@healthdatamovers.com to get started!