Modern healthcare service records, called Claims, record the medical treatments by a Provider (Doctor/Clinic), medication advised etc., along with the charges, and payments to be made by the patient and the Payer (insurance provider). Denial and rejection of healthcare claims is a significant administrative burden and source of loss to various healthcare providers and payers as well. Automating the identification of Claims prone to denial by reason, source, cause and other deciding factors is critical to lowering this burden of rework. We present classification methods based on Machine Learning (ML) to fully automate identification of such claims prone to rejection or denial with high accuracy, investigate the reasons for claims denial and recommend methods to engineer features using Claim Adjustment Reason Codes (CARC) as features with high Information Gain. The ML engine reported is first of its kind in Claims risk identification and represents a novel, significant enhancement to the state of practice of using ML for automating and containing claims denial risks.