Austin, Texas-based firm also discovered that as companies diversify their usage of AI, previous methods of cybersecurity fail to address prevalent risks.