Autocoding 2.0 Overview

What is Autocoding 2.0?

Autocoding 2.0 is HANK's AI-powered claim processing system got an upgrade. It is now built on a Large Language Model which allows for scalability, ability to mark cases as Code Complete automatically, and increased configuration options. It does this through the newly introduced Human Review Elements (HREs) ecosystem, which act as targeted quality checkpoints. Each HRE tells you:

• What needs review
• Why it matters
• Who should fix it
• What action to take

HREs are organized into four priority levels:

• HARDSTOP – Must be resolved before the claim can be submitted
• FINANCIAL – Won't block submission, but affects reimbursement
• COMPLIANCE – Regulatory/audit alerts that protect you during reviews
• INFORMATIONAL – Awareness-only notices, no action required

How to Enable & Use It

Reach out to [email protected] to see if your organization is running Autocoding 1.0 or 2.0. Enabling Autocoding 2.0 requires special coordination with the HANK AI staff.

Once Autocoding 2.0 is enabled…

• Turn on the work queue in Claim Maker
• Filter by your role to see only HREs assigned to you
• Review each flagged item and resolve, document, or escalate as needed
• Mark items complete and move to the next encounter in your queue

Key things to know:

• Autocoding only runs once per case — if HREs prevent full autocoding, a human must manually apply codes
• HARDSTOP HREs block submission until resolved *note- you MAY still receive autocoding predictions even if it has HARDSTOP HREs.
• Your facility can customize confidence thresholds, toggle specific checks on/off, and set code exemptions by reaching out to [email protected]
• For issues or feedback, contact [email protected]