Robotic process automation (RPA) lets businesses quickly transition analog, human-led activities into automated digital actions.
What differentiates RPA from tools that have complementary or overlapping capabilities is the ability to implement computer-initiated actions that would typically require human intervention. Examples of such actions include pulling information from an unstructured document, using a visual interface to enter or query data, extract information from printed forms, or evaluate voice input. Without this capability, digital automation is not possible and, therefore, neither are the benefits associated with such automation.
RPA is democratizing automation—empowering employees and increasing efficiency within organizations. But the implementation of RPA processes within a business can be difficult—and often there are alternatives that are more cost-effective and better aligned with business needs.
In this report, we lay out the questions and considerations an organization needs to examine before choosing RPA integration. We look at:
- Situations best suited to RPA deployment
- Where alternative solutions might be considered
- How RPA compares to low-code tools
- Staff expertise
Can RPA Support the Entire Enterprise, Or Only Front-Office Automation?
RPA has an advantage through visual and low-code development capabilities that enables citizen developers and non-IT staff to automate routine front-office tasks. And IT teams can easily write automation functions for back-office integration in high-level programming languages. Moreover, today’s serverless platforms and PaaS vendors allow developers to focus on the specific integration task while leveraging cloud services.
This means IT can respond very quickly to basic API-based application integration tasks, leveraging low-cost platforms and services. In contrast, most RPA platforms incur additional license overhead on a per-bot basis to scale and operate bots in parallel.
Back-office integration will typically see transaction ratios of 30:1 or higher when compared to front-office integration, so coding the integration becomes much more economical.
What Problems Could RPA Solve?
Here are some of the key use cases for RPA:
Assisted Data Capture – RPA tools can review the information being entered into a form in real time and assess for accuracy. They can also pull information from other systems to fill in fields, and enable new mobile and voice-based systems by taking care of detailed form entry within the bot.
Assisted Customer Service – RPA tools can monitor message boards, email boxes, social media, and other sources where customers may be seeking assistance or raising concerns about products and services. This capability is significantly enhanced in those that have democratized AI with access to natural language processing facilities to go beyond common terms and infer context.
Business Process Outsourcing (BPO) – Many RPA tools can help analyze multiple input sources and either immediately address the need, for example to pay a claim, or place the request into an exception-handling queue to be addressed by a human. This significantly reduces the number of humans required to process large volumes of transactions daily.
Can RPA Tools Replace iPaaS and Associated Integration Tools?
RPA can automate many functions, but if the goal is significant data remapping and restructuring, iPaaS tools will provide a better overall solution.
iPaaS platforms tend to be licensed on a server basis whereas RPA tools are usually e licensed on a per-bot basis, allowing iPaaS tools to scale in a more cost-effective manner than RPA.
However, if transaction volumes are low, or if the automation is a combination of both front-office and back-office tasks—for example, pulling data off the web and updating a CRM or ERP application via an API—the business may benefit from the simplicity of having RPA complete the entire task.
Can RPA Tools Provide Low-Code Capabilities?
Today’s RPA tools often overlap with low-code tools; however, RPA tools typically do not address user experience and are used to create headless services or “bots,” whereas low-code tools focus on web and mobile applications development.
Some RPA tools allow you to augment the visual workflow with traditional programming. This capability can be very useful for handling some complex tasks. However, it also opens a can of worms regarding long-term support for the bot.
If you need to build a complex task, your business would best be served by developing it as a separate microservice and then calling the microservice from within the RPA application. For some businesses, this may lessen the value of using an RPA solution rather than IT to alleviate the backlog of automation work.
Can Automation Tools From Amazon and Microsoft be Used in Place of RPA?
Cloud service providers, such as Amazon and Microsoft, have a diverse set of services available to developers that can be provisioned and operated through a set of programming interfaces. This makes them very accessible and easy to combine to deliver automated business processes.
Moreover, there are products, such as PowerApps and Flow from Microsoft, that democratize the use of powerful services like natural language processing, image recognition, optical character recognition, and database access through visual modeling. While these tools cannot automate existing user interfaces, they do offer usable alternatives to using RPA to create unattended bots for routine tasks, such as responding to email or searching social media and the web.
What Pricing Model is Most Appropriate?
While SaaS pay-as-you-go might offer better overall pricing, it may not be suitable for working with an organization’s internal applications without expensive connectivity options.
Pricing models vary widely when it comes to RPA products. Most have support for deployment in owned data centers and as SaaS. They also offer licensed runtime capabilities and pay-as-you-go subscriptions.
Organizations need to consider how the pricing model will affect their ability to parallelize certain bots if they want to scale.
Many RPA tools have some limits they place on execution of bots, and often the schedulers enforce serialized execution. Overcoming this limitation usually requires additional licensing.
Businesses may also be constrained by the need for integration with other software applications.
What Should I Worry About Regarding Security?
The potential for leakage of confidential information or enabling systems to be breached more easily increases significantly with the introduction of the citizen developer. To combat this, we advise creating a center of excellence (CoE).
One of the key tasks of a CoE is to ensure bots adhere to governance related to security. Here are some factors a CoE should include in its governance requirements:
- No embedded credentials
- Proper use of privileged access management (PAM) and vaults
- Minimize risk of unauthorized users accessing a system in attended mode
- Ensure bot design does not interfere with system operations
Are There Enough Skilled Individuals to Support the Rapid Growth of RPA?
Due to vendors’ investments in publicly available software and education, many system administrators and programmers have been trained to use RPA tools. However, interest in automation is growing rapidly and there will be continued need for those who can manage the RPA environment, build centers of excellence, and drive governance around the use of RPA in the enterprise, as well as those that can help companies evaluate existing processes and incorporate RPA where appropriate.
Introducing RPA into a business is not simply a matter of technical implementation; it is a strategic and business-focused process that centers on an organization’s needs, targets and budgets.
RPA can make a big difference to organizations large and small, but it needs to be aligned with the strategic goals of each business. IT leaders must be sure that its implementation will bring business benefits and ROI that could not have been achieved otherwise with more easily accessible or cheaper tools and processes.
RPA streamlines repetitive tasks, freeing up employee time and increasing productivity—but this can only be achieved if it is implemented for the right processes and if you have trained staff who can successfully manage and incorporate it correctly.