Robotic process automation (RPA) is an umbrella term for a group of technologies that enable line-of-business users to set up, launch and administer virtual workers without relying on an IT department. RPA tools can link different applications together and automate routine tasks so they don't require direct human intervention. The general methodology for an RPA deployment is two-fold: There's a user interface allowing end users to craft automated workflows and a suite of scripts that automate routine and non-varying data input or data processing work. RPA platforms are increasingly being augmented with machine learning and with the idea that the "robots" will be able to iterate routines and optimize how they complete their data processes.
RPA is a small but fast-growing software market. Gartner estimated the segment’s worldwide revenue at approximately $850 million in 2018, representing a 63% year-over-year growth rate. RPA is appealing from a cost perspective because it frees up workers to do more complex and less routine work, reduces overall labor costs, and frees up IT professionals so they're spending less time implementing and supporting middleware for their colleagues. It also appeals to business users who see it as a way to bypass the usual IT processes of acquiring and integrating new technologies into an enterprise's current legacy stack — RPA platforms do all the process automation and data integration for them.
However, RPA has a few drawbacks. The ease with which it can be implemented to automate work depends in no small part on whether the organization's data is easily assimilated into any automation models.
Last year, Ted Friedman, distinguished VP analyst for Gartner, said, “As organizations accelerate their digital business efforts, poor data quality is a major contributor to a crisis in information trust and business value, negatively impacting financial performance.” It also impacts overall operations and machine learning. According to Alegion, a training data provider for AI and ML initiatives, a majority of enterprise organizations trying to use machine learning at work have run into problems with data quality, the data labeling required to train AI and building model confidence.
If an organization has a vast trove of unstructured data, such as handwritten data, images or audio content, it requires more complicated and expensive AI-based technology to automate the process efficiently.
The other big drawback: Most of the low-hanging automation fruit — structured and repetitive tasks — has already been automated or is being automated through scripting, macros or simple RPA deployments. Consequently, RPA could include significant planning and implementation costs, because what's left for an enterprise to augment via RPA is much more complex than earlier automation tasks.
Our list of the top 10 RPA platforms has been selected using criteria to winnow down entrants in the RPA space to those that best met the following criteria:
Does the vendor offer a programmable bot solution? Being able to customize and refine robots based on an individual customer's business processes was a must-have for our list.
Is the vendor enhancing its automation tools and processes with machine learning or other cognitive tools? Having bots that can correct themselves and adapt their routines based on learning from the data is critical. It's also future-proofing any RPA platform offering; natural language processing and image recognition are going to be must-haves soon, if they’re not already.
Does the vendor have a wide range of experiences across a wide range of industries? Vendors that have successfully figured out how to make their services and tools work in different industries will have a better likelihood of being able to make an RPA implementation work. They'll also be more likely to have internal processes in place to figure out how to assess a customer from a new industry vertical and see where RPA works with that industry's needs.
Does the vendor have an active support plan for customers and an active user/vendor community? Although the appeal of RPA is the set-em-and-forget-em model of automating work and then lightly monitoring the robots, there is something to be said for being able to turn to fellow customers or vendors in the same space to see what's possible in deploying and optimizing RPA platforms as offered via a specific vendor.
Are customers satisfied with the deployment and support the vendor provides? A comparatively smooth implementation means less business downtime for the customer and faster implementation of RPA benefits for the user.
Is there a central dashboard allowing customers to monitor the workflow metrics and thus assess operational performance? There's no point in deploying RPA to optimize the enterprise if there's no transparency into how the deployment is performing.
Without further ado, here’s the list.
Founded in 2003, Tethys Solutions rebranded itself as Automation Anywhere in 2010. Since then, it's developed a product portfolio including Bot Insight, an analytics platform that delivers real-time operational and business intelligence; BotFarm, an enterprise-grade platform for scaling bots on demand; Bot Store, with pre-built bots for business automation processes; and an RPA mobile app. The real sell for clients: an array of integration components that users can string together to create automation scripts. There are also components that allow Automation Anywhere to provide extensions to its product that allow users to access other third-party services they may already be using, such as IBM's business process management tools. Automation Anywhere has considerable geographic reach, with more than 1,600 enterprise clients and customer relationships including Dell, General Motors, Tesco, AT&T, Whirlpool, Wipro and LinkedIn.
This U.K.-based enterprise software and services company was one of the first vendors to recognize that task automation could be turned into a huge market for products and services. Its Intelligent RPA Platform is composed of three elements: Object Studio, which allows end users to build process automations via a drag-and-drop interface and to reuse the building blocks of those processes (i.e., the action triggers and follow-up actions); Digital Workforce, autonomous software robots that have automation-centric AI skills; and a Control Room, which assigns processes to the robot workforce and manages workforce demands, scheduling and optimization. Blue Prism’s RPA platform can be implemented on-premises or via SaaS. The company is also keen on compliance; it cites activity logs that create an evidence-based audit trail of every system event, which can be used when required to show proof of compliance.
The professional services company offers a set of RPA-driven services under the umbrella of its Intelligent Process Automation Solutions. Among its RPA wins across different industries: Health care claims-processing bots process 3.8 million transactions annually for insurance firms; invoice-processing bots process over 1.2 million invoices annually for retailers, reducing delayed-payment interest charges by more than 50%; and exchange-derivative-clearing bots process over $3 billion annually for commodity futures trades. The company also offers robotics as a service (RaaS). In this scenario, the robots are deployed on virtual machines alongside applications they interface with and take over easily routinized tasks, freeing human workers to do the less routine tasks or work through anomalous situations.
This wholly owned Infosys subsidiary offers AssistEdge Robotic Process Automation, an end-to-end service for building and implementing RPA. Its latest version, AssistEdge RPA 18.0, introduced a cognitive engine dubbed Albie that learns from a variety of data sources to deliver insights across process design, management and execution. The RPA offerings provide integrated software robots to automate high-touch, repetitive processes, plus assisted automation for processes that need manual intervention. There is also a community version free for anyone to experiment with writing processes for their own personal robot. EdgeVerve has also recently debuted a call center RPA product, AssistEdge Engage. The company says, "AssistEdge Engage combines rich capabilities of AI and automation, improving agent productivity by automating repetitive tasks, cross-sell and upsell by offering rich customer context."
This 21-year-old business service provider focuses on automating processes in industries like banking, insurance and health care. Its product portfolio centers around an artificial intelligence platform called Amelia, and on this platform, IPSoft has built Amelia City and Amelia Marketplace, which provide off-the-shelf options for automating processes. The company also offers RPA products for IT management through IPCenter, a library of over 20,000 end-to-end automations. IPSoft says its customers see a 35% gain in efficiency in the first 90 days after deployment.
The company's flagship RPA product was recently renamed from Kofax Kapow to Kofax RPA. Because of a strong, pre-existing customer base in image capture services, there's a compelling combination of RPA tools and document capture. For many customers, it would be less expensive to buy a bundled Kofax RPA and OCR offering than to use an RPA service that would require separate OCR licensing. Kofax's RPA offering works on a slightly different model than that of other RPA platforms: By running a small agent on the desktop, which then communicates with robots on a central server, Kofax RPA contains its robots on one virtual machine that is centrally managed. There is no need for the staff to update individual robots or to push them out to individual machines. Not only does this reduce IT overhead in man-hours, but it also lowers the cost of the associated infrastructure.
Founded in 1983, this enterprise software vendor is making a play in the RPA space by tightly integrating its RPA offerings with its existing CRM and business process automation products in addition to offering a standalone product. The company is aiming at IT departments that have concerns about legacy technologies and how RPA platforms work with them, writing of its RPA offerings, "Pega Robotic Process Automation is non-invasive and equips organizations to bridge legacy systems, easily close data integration gaps and wrap legacy system integrations without making changes to underlying technology investments." Among the products it offers, Pega Robotic Desktop Automation is aimed at automating the tasks and multiple-application workflows associated with customer service representative workloads. The Pega Robotic Automation Studio, meanwhile, allows customers to create and record RPA workflows for the robots while Pega Robotic Automation Console manages the work, SLA compliance and processing activity for Pega software robots via dashboards and reports.
The Netherlands-based company focuses its RPA efforts on reducing manual labor across the supply chain, logistics, forecasting, e-commerce delivery and financial posting. Redwood Robotics software includes ready-to-use automated process activities for all of those areas. Those automated tasks employ standard connectors that work directly with business suites like Microsoft Office and enterprise applications like SAP, Oracle and PeopleSoft. Redwood Robotics also offers a build-your-own automated process studio and a dashboard for viewing processes in four states: currently scheduled, in progress, awaiting review and recently completed. The company also stresses its ability to integrate control and compliance capabilities into the automated processes and automatically publishes process audit information to an accessible yet secure archive available to both internal and external auditors.
Another of the top RPA platforms is UiPath, which designs and develops specialized RPA software. The company offers its customers an automation design tool, and it supports both on-premises and automated cloud deployment. UIPath has over 100 technology partners that offer complementary technology and tools, so it supports integrations with major products and applications covering business process management, process mining and AI. Its product line includes the following: UiPath Platform, which delivers rapid automation of manual, rules-based and repetitive processes; UiPath Studio, which is a drag-and-drop, code-free studio editor that allows customers without any programming experience to access an extensive library of activities and pre-built automation components to build out automation workflows; and UiPath Robots, which execute the automation workflow designed in UiPath Studio and can be centrally managed from UiPath Orchestrator, the centralized robot management dashboard where you can deploy, secure and manage UiPath Robots at scale.
Founded in 2010, WorkFusion is aiming for the global operations market by offering the one-two combination of business process management and robotic process management. The company says one of its largest differentiators is a unified platform, so customers have one central admin tool to automate processes and manage specific roles for users. WorkFusion recently announced a partnership with NEC, where the company collaborated with NEC Management Partner (NMP), NEC's shared services organization, to develop expert cognitive bots that address specific finance and accounting processes that are unique to Japan. WorkFusion uses machine learning to power its software robots for knowledge work. Its customers include those in the banking and financial services, insurance, healthcare, consumer products, utilities, telecom and retail industries. In 2018, the Alconics — an independently judged AI awards program operated by Informa — awarded the Best Innovation in RPA award to WorkFusion Smart Process Automation (SPA) for letting its customers automate complex processes from end to end on a single platform. In another awards ceremony, TechXLR8 Asia judges selected WorkFusion SPA as the Best Application of AI in Financial Services for automating complex processes with unstructured data such as claims handling, using uniquely small training data sets of only 400 to 500 samples.