The Company

WorkFusion was founded by Andrew Volkov, Mikhail Abramchyk, and Max Yankelevich in New York in 2009. Originally named CrowdComputing Systems, with a focus on microtasking crowd-computing work, it changed the company name to WorkFusion in 2014 to better reflect its broader applicability of task and work automation. Today, the company is led by CEO Adam Famularo, who joined the firm in September 2021. WorkFusion has raised $341.3 million and is currently at Series F, with this last round of funding in March 2021.

The Technology

The WorkFusion product reviewed in this report represents a fascinating evolution from a focus on crowd computing through robotic process automation (RPA) to digital workers. So, at the top line, WorkFusion today has a platform that provides digital workers. But rather than RPA bots undertaking simple, repetitive tasks (lift and shift address data from application A to B, etc.), WorkFusion delivers highly specialized and named avatar-like “digital workers” to undertake specific but often complex work routines. For example, Ilana, the insurance underwriter, or Darryl, to review documents submitted to a bank by its customers, or Evelyn, to monitor negative news (see Figure 1). It’s also possible to build and deploy custom digital workers on the platform. These digital workers essentially augment, assist, and automate critical elements of human work.

The question is, how? Although WorkFusion can typically be described as an RPA vendor, RPA is only a small part of the story. In truth, WorkFusion is hard to categorize as its approach and roots are pretty distinct from the perceived market competition. To understand how WorkFusion works, it’s best to start at the bottom and move up the stack.

Core to WorkFusion’s approach is the company’s roots, human-in-the-loop technology for training AI. Indeed, the company was a pioneer in building a platform to distribute and collate microtasks to crowd workers. While “crowds” are no longer part of the approach, operations analysts at banks and other organizations still play a role in working side-by-side with the digital workers. So, for example, if the digital workers are unsure whether a mark on a document represents the letter I or the letter L, that element can be part of a “manual task” sent to an analyst or operator for human identification and verification.

WorkFusion’s platform is essentially a low-code/no-code workflow and machine learning development environment that comes with pre-trained, ready-to-deploy AI models and, of course, analytics and dashboarding capabilities. Important to note here is that WorkFusion’s software provides some advanced intelligent document processing (IDP) functionality in the form of AI-driven OCR and NLP, or in popular parlance, “document understanding.” These IDP tools can scan and read anything from a driver’s license to a contract, a logo to a barcode. In addition, it provides a range of industry-specific IDP models. As it is AI-based, it will, with human intervention, improve over time rather than degrade as traditional document capture systems do. Finally, there are rules-based mechanisms with WorkFusion Enterprise (RPA) that are triggered to automate processes and task activities. In short, there is a lot here, and from a technology standpoint, it’s good, goes beyond most of the competition, and has an extensive range of potential use cases.

In practical terms, WorkFusion extracts, reads, understands, and processes data. It can extract data from structured or unstructured document sources and then enrich, validate, and classify the data. More interestingly, the AI modules can interpret sentiment from the captured data and, if required, write summaries and responses and compile groups of related and linked data sets. This enables it to apply quite complex decision-making rules and automate workflows. WorkFusion is available to run in a private cloud (or on premises), as a managed service, or via an SaaS model. It can also integrate into third-party systems via a REST API or pre-built connectors to applications such as Microsoft Dynamics and Salesforce.