Robotic process automation (RPA) technology deserves a place in your automation toolkit, but it probably shouldn’t be the only tool you use. In this article Neal Gottsacker Chief Product Officer of Nintex explains how workflows, digital forms, and document automation also have important roles to play in enterprise process automation.
Who doesn’t like the idea of handing off tedious busy work to a reliable, uncomplaining digital assistant? That’s the promise of robotic process automation (RPA) technology. The promise is compelling enough that it’s driven a dramatic surge in RPA adoption over the last several years, allowing its providers to establish huge valuations and attract massive amounts of investment capital.
But there is more to the RPA story. It is indeed a useful technology, but it’s not a panacea. Enterprises considering whether to adopt it should have their eyes wide open.
First, some basics. RPA “bots†mimic human keystrokes and mouse clicks. Anything a human can do on a screen, a bot can do too – faster and more reliably than a human can, without taking breaks or making mistakes. Bots can be trained to log into applications, copy data from one location and paste it in another, copy files, harvest data from the web. Bots can open email and attachments, fill in forms, read from and write to databases, make calculations, collect social media statistics, extract data from documents, and more, all very quickly and reliably.
RPA use cases fall into two main categories: attended and unattended. In attended use cases, humans initiate and control robot tasks, often by embedding bot functions within apps. Think of a front-office employee interacting with multiple desktop apps – email, Excel, customer records, and the like — in a role such as claims processing. By contrast, in unattended use cases, bots completely replace humans, often serving as a virtual workforce to do batch jobs such as rapidly updating thousands of bank records.
Learn More: RPA: A Must-Have for Digital TransformationOpens a new window
The speed and reliability of RPA bots is driving their adoption in multiple industries, including financial services, healthcare, IT, manufacturing, retail, telecoms, transportation, and utilities. In both attended and unattended cases, RPA bots are effective because they execute the process in exactly the same way, time after time, without mistakes.
So when they’re running smoothly, life is good. But here’s the thing: They aren’t always easy to set up, and they don’t always run smoothly. For setup, many RPA bots require developers to write software code. This takes time and money, which is why most of the major RPA vendors have established partnerships with technology systems integrators. It’s a marriage made in heaven. Also, bots can be fragile. Because they can only follow human/user interface paths – UI paths – they can’t adapt when UIs change, which happens frequently. With every UI change, each bot must be adjusted.
Perhaps most important, however, is that RPA bots are very literal. They don’t exercise judgment or make decisions. But judgment and decisions are fundamental to many enterprise processes. Think about sales operations: Should you grant a discount to a favored customer? How deep a discount? Think about communications: Has everyone involved in issuing that press release reviewed it, and submitted comments or revisions? Think about human resources: Has everyone reviewed the latest batch of applicants and given their comments, so you can rank them all and decide which to hire? These are just a handful of the thousands of processes in a modern business that call for collaboration, judgment, decisions, even creative thinking. RPA bots can’t do this kind of thing.
But you wouldn’t know that from listening to the RPA vendors. They have grown their businesses and valuations in part by explaining what their bots are good at, and in part by making RPA synonymous with automation itself – that is, by describing every enterprise challenge as one that bots can solve. They tend to steer the discussion away from topics like judgment and decisions. One result is that, by some estimates, one-third to one-half of initial RPA deployments fail. Another is that scale remains a major hurdle for enterprise RPA: According to Forrester Research, more than half of enterprise deployments have fewer than 10 bots, and less than 5% of enterprise deployments have more than 200 bots.
Learn More: Integrated Automation Platforms Can Fuel Businesses, DevelopersOpens a new window
This is not to say that RPA doesn’t deliver real business value. It does – when it is properly understood and correctly deployed. To succeed with RPA, it helps to understand that RPA is just one form of automation. There are several others, including workflows, document automation, and digital forms. Workflows are well suited for processes that call for collaboration, judgment and decisions, as described above. Document automation is extremely useful for easily creating customized proposals, contracts, and correspondence. Digital forms can greatly accelerate and minimize errors in processes currently executed using paper.
As this discussion suggests, there are various types of automation – because there are many types of processes, and each may have multiple different elements. Here’s a hypothetical example that exists in similar form across dozens of industries: One step might call for cutting and pasting from a web form or a spreadsheet, something ideal for RPA. Another step might involve entering data into a form – and digital forms are so much more powerful than paper, with capabilities such as dropdown lists, spell-check, instant transmission via smartphone, automatic upload to a system of record. The next step might call for review/approval, which is well suited for a workflow. If the goal is to generate a custom sales proposal, that could call for automated document generation. The document might need to be signed by multiple people, something that legally binding e-signatures handle rapidly and easily.
So how do you know which type of automation to use? Start by mapping your processes. Use a visual tool that lets you describe them in plain language, identify what types of information and systems of record are involved, which people need to participate, what decisions need to be made. Share the process map with the relevant people in your organization to get feedback and buy-in – often, at this stage, you’ll get suggestions that help you improve the process better right away.
Learn More: Creating Business Value with Intelligent AutomationOpens a new window
Now, with an agreed-on process map, you can make intelligent decisions about what type of automation to apply – RPA, workflows, document automation, digital forms, whatever is right for your specific process. Then you’re in a position to start evaluating automation tools. You’ll want some breadth of capability, so you can use the right tool for the job — apply the right type of automation to the process at hand, according to your map. You should consider ease of use, because automation solutions are not created equal in this regard. Some require extensive code, some require none at all – they are completely no-code. Similarly, some solutions require extensive professional services, while others require very little.
Once you’ve selected your automation tools, how quickly will you get results? It’s realistic to think in terms of hours or days; you shouldn’t have to wait weeks, months, or years to get real business value. And how will you know? Some benefits will quickly be evident. For example, you might get e-signatures on a contract in minutes, when it used to take days to print and mail paper copies. Or perhaps, at your bank, you used to employ a team of people to update a set of customer records each night, but now your RPA bots now do the job almost instantly.
For fine-tuning your processes, consider an instrumentation and analytics solution that can show you when, where, how often, and how quickly your processes are running. With that insight, you can quickly see and address bottlenecks and breakdowns.
The key to success is to create a foundation by mapping the processes you’re trying to automate. Then you can select the tools that fit best. RPA may well be among them, but it likely won’t be your only tool.
Let us know if you liked this article on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We would love to hear from you!