The Company

FortressIQ was founded in 2017 by current CEO Pankaj Chowdhry. The firm is headquartered in San Francisco and is currently at Series B funding, having raised $46 million. FortressIQ is backed by Lightspeed Venture Partners, Boldstart Ventures, Comcast Ventures, Eniac Ventures, M12, and Tiger Global. The company has around 70 employees.

The Technology

Process mining/intelligence has come into vogue over the past couple of years; indeed, the automated analysis of business process activities holds a great deal of promise and potential for future growth. However, it involves many different layers, techniques, and requirements, and in parallel many different activities and sources of data that can be analyzed. As a result, no two systems will approach the challenge in the same way, produce the same results, or meet the same needs.

FortressIQ has a distinct and interesting angle, delivering what it terms a “Virtual Process Analyst” that can be deployed either on-premises or via a virtual private cloud. The VPA works by embedding a sensor in a user’s desktop environment, visually recording the activities, then analyzing video of the user’s screen. At first glance, then, VPA seems like a form of advanced screen scraping, but a look under the covers reveals something much more sophisticated.

FortressIQ makes use of computer vision (CV) along with various machine learning (ML) and artificial intelligence (AI) techniques. Its CV model has been trained to understand all the different controls that may exist on a screen, from checkboxes to drop-down menus, and the data that each contains, etc. The CV runs in parallel on multiple users’ screens, assuming they are all doing the same kind of work. This way, the CV is able to understand each worker’s variations and differences in approach, and, equally importantly, their screens’ different layouts and configurations. Collectively, these multiple screen video captures can then be normalized across natural variations, with any unnecessary or irrelevant actions and noise filtered out (for example, checking personal websites or mail), thus cleaning the data. Finally, a unified (across all applications) log file is created. This file provides the basis for an accurate and nuanced “as-is” picture for analysis and process optimization.

Such a file, composed as it is from observing the actions on a user interface, needs to be managed carefully. First of all, it must be filtered to ensure that compliance requirements for personally identifiable information (PII) are met, as well as any broader privacy compliance needs. FortressIQ uses what they term a “Privacy Enhanced Gateway” (PEG) to filter and automatically redact sensitive data. The PEG is essentially a pre-trained AI-based, CV firewall. Any data collected from the screens is filtered, and sensitive data removed/redacted, through the PEG in a virtual private cloud environment before further analysis. The FortressIQ PEG system is compliant with HIPAA, Information Security Management 27001, and AICPA Soc 2.

This novel data-collection and data-cleaning process is at the heart of FortressIQ – the output is initially processed with process-mining algorithms and neural network sequence models in addition to the natural language processing (NLP) models, then displayed in a dashboard-style user interface. This provides accurate insights for human business analysis into what is really happening within or across business applications and processes. FortressIQ is, then, an automated business analysis tool, automating to some extent, but rather augmenting, the real-world analysis and optimization that will be undertaken by human experts reading and using visualizations to optimize outcomes. Clearly such a tool has many uses: for example, identifying potential RPA opportunities to improve customer or employee experiences. Finally, it is worth noting that FortressIQ can be integrated with Microsoft Power Automate.