When possible, Claim Maker uses the extracted data from your files to make line item coding predictions. Predictions can include:

- CPT procedure codes: surgical, anesthesia (ASA), radiology, and interventional radiology
- ICD-10 diagnosis codes
- Modifiers
- E/M visit levels for office, hospital, and emergency department visits. The system selects each level with deterministic rules applied to evidence it cites from your documentation.
- Surgeon companion claim line items, for organizations that enable companion claims

Predictions appear at the top of the line item. The coder can apply a code individually by clicking on it, or select **Apply All** to apply all predicted codes.

Hover over a prediction, or over an applied predicted code, to see the context for why that code was chosen, together with a confidence score. High-confidence scores appear green. Lower scores appear yellow and should be reviewed. Your organization can customize these confidence thresholds. Contact ui-support@hank.ai to learn more.

![Hovering over a prediction shows the context for why that code was chosen and its confidence score.](/uploads/2026/06/136733d8-5ae0-485e-ad88-73a3ac9063cd.png)

Autocoding predictions also provide the reference document the predicted code came from. A hyperlinked document appears in the predicted code pop-up. Click it to reveal the source document, then click the document to navigate there.

Inside the document, the **Summary** button generates a summary of the information used for a specific code.

![The Summary button inside the source document generates a summary of the information used for a code.](/uploads/2026/06/0663d81e-ef49-49c4-9fa0-c14eb8f4607e.png)