A Hybrid Cognitive Contract Application for Identifying Accounting Risks in Contractual Language

Abstract

Analysis of legal and contract documents often requires the discovery of document structure and important elements such as party (buyer, supplier), nature (obligation, right) and category (warranties, delivery). Particularly, the process of reviewing and assessing potential accounting risks in contracts is critical yet typically manual driven that in- creases the risk and level of inaccuracy, which also translates on additional time investment and secondary controls to be executed by accounting to ensure revenue is properly recognized. To mitigate this problem, we developed an end-to-end system Cognitive Contract Application (CCA) to assist users to automatically assess and identify accounting risks in contracts. Our system makes a great impact on the contract review pro- cess of our organization in two-fold: i) reducing significant amount of manual works from users, and ii) recovering risks worth millions of dollars missed by the legacy manual contract review process.