On December 3rd 2019 Gartner published its "Critical Capabilities for Robotic Process Automation" report. Here are the most important takeaways:
- With all the hype surrounding the Robotic Process Automation (RPA) market, most companies buy software from one of the top 3 vendors; however, there are other strong, competitive products on the market that offer innovative features to customers.
- Gartner has identified 10 key areas of functionalities associated with RPA, which they have divided into 3 categories: "core" (must haves), "critical" (should be available) and "tangential" (adjacent to and often used in conjunction with RPA).
- In this fast moving, competitive market environment, there was little differentiation between the weighted scores of the top 10 products in the use cases Gartner investigated. As vendors battle for the top positions and try to outperform each other's offerings, a mini arms race is taking place in each area of the functionalities. Most established vendors are gradually expanding the scope of their product offerings and closing more and more gaps in their tangential functionality coverage through partnerships with third-party vendors.
Enterprise architecture and technological innovation leaders evaluating RPA tools and techniques for technology-based digital transformation should pay attention to the following:
- Explore the relative merits of different products on the market by distinguishing between the needs of end users and those needed for effective reuse/scaling across the enterprise.
- Optimize the likelihood of a successful RPA initiative by pre-defining the enterprise use cases and key business functions. Alignment with strategic goals and expected business results will help you choose the right product in the long term.
- Maximize ROI by prioritizing respectively considering a RPA suite with advanced functionalities. This allows you to better support the company's primary use cases. A close examination of the capabilities and limitations of the tangential functionalities, such as those of broader process automation and machine learning, will also allow you to look beyond the initial proof of concept and estimate the potential sustainable value of the RPA investment.
Innovation leaders should use RPA to automate predictable, rule-based tasks by mimicking the way people interact with the user interface (UI) of the various applications they use. Thus, RPA is a tactical tool that is most effective when applications are used in a manual task today and have either no or incomplete APIs. RPA, on the other hand, is not suitable for the automation of rapidly evolving or poorly understood business processes.
The Gartner Report
Gartner's analysis evaluates vendors and their solutions based on product functionality only. The research team provided 3 examples of use cases to the vendors and then based their ratings, rankings and weightings on what the vendors had demonstrated as results. Prior to and during the product demonstrations, the analysts clearly communicated what the relevant focus areas are.
You can understand this analysis as a supplement to the "Magic Quadrant for Robotic Process Automation Software" report. The Qudrant highlights a number of factors. These include vendor profitability, vision, their marketing and the geographic focus of the various vendors.
The 10 RPA-critical functionalities
Gartner has identified 10 RPA-critical functionalities that they use in their analysis to help differentiate between the most popular products on the market. As mentioned at the beginning, they have divided these functionalities into 3 major groups:
- Core: Automation development, integration features and control panel/dashboard. These core competencies are absolutely central to the operation of an RPA product.
- Critical: Here the focus is on component/script library, security, resilience and error recovery as well as the detection and handling of external change influences and drivers from the environment.
- Tangential: Artificial Intelligence (AI)/Machine Learning (ML)/Natural Language Processing (NLP), alongside broader business rules/process automation and finally Optical Character Recognition (OCR). These 3 competence groups are located in adjacent technology areas that a vendor may have developed or could provide in a partnership.
For each defined functionality, the analyst team identified a number of supporting functional elements for which they developed a specific evaluation spectrum and procedures. They found that the best results for an individual functionality are achieved when the vendor develops it itself and delivers it to customers as part of its own RPA tool (without additional licensing). If the functionality was purchased externally and/or additional licensing was required, the team had to reduce the functionality score from the highest achievable level in the assessment spectrum to a lower level.
The 3 most common use cases of RPA
Gartner also identified the 3 most common use cases for RPA to filter and prioritize features for those that have the highest level of impact:
- Integration via UI. This case focuses on the "unattended" use of RPA, which involves moving or copying a small amount of data between two applications simply and relatively ad hoc.
- Large-Scale Data Migration. This is actually about integrating several cooperating component sets to move data "en masse" between two different applications. The functions required for this use case are similar to those needed to build an "API facade" - the effective reuse of discrete RPA automation for the last mile of integration.
- Augment Knowledge Workers. This case focuses primarily on the "attended" scenario where RPA is used to enhance the ability and capabilities of knowledge workers to do their jobs. This usually saves them time or gives them more capacity to take better care of their customers, as the RPA robots take care of the heavy lifting of the information assembly.
Results of the study
The resulting findings can be seen below in a summarized list of product scores, categorized according to the 3 use cases just described:
The scores range from 1.0 to 5.0 and should be interpreted as follows:
1 = Poor or non-existent: The majority of the defined requirements for the functionalities are not covered
2 = OK: Some requirements are not covered
3 = Good: Meets the requirements
4 = Excellent: Meets or exceeds some of the requirements
5 = Outstanding: Exceeds the requirements significantly