Developing a Cognitive Assistant for the Audit Plan Brainstorming Session

With technological advances, audit firms have been able to increase the level of decision support embedded within the firms’ audit support tools (Dowling et al., 2008). Large audit firms have been investing substantive resources into the utilization of Artificial Intelligence (AI) to take advantage of their past audit experience and industry knowledge (Kokina and Davenport, 2017; Appelbaum, 2017). During the initial stages of an audit, the engagement team will meet to update their understanding of the client through brainstorming sessions. Audit firms are required by SAS No. 99 and SAS No. 109 to use brainstorming sessions to evaluate risk factors and discuss the susceptibility of the entity’s financial statements to material misstatement, either as a result of error or fraud (AICPA, 2012). At present, the most commonly used decision support tool for audit plan brainstorming is the checklist (Bellovary and Johnstone, 2007), which has shown limitations (Dowling and Leech, 2007; Seow, 2011; Landis, 2008). To improve audit plan effectiveness, this study proposes an audit domain cognitive assistant system to provide interactive decision support for information retrieval and risk assessment in audit brainstorming session. Cognitive Assistants are speech-enabled technologies that can understand voice commands, recognize conversation’s context, and answer questions in a personable manner (Garrido et al., 2010; Myers et al., 2007)