The pressure to automate is omnipresent. Every organization wants to eliminate manual, sit-at-the-computer tasks.
In a 2018 Deloitte global CIO survey of 1,437 executives, 96 percent of respondents considered digital a strategic priority. When asked about the primary focus area of digital within their organizations, 69 percent said process automation and transformation were #1 of importance.
But the prospect of automation can be daunting. Between identifying use-cases and prioritizing approaches such as no-/low-code, open source, and robotic process automation, you might not know where to start.
In this blog, I suggest you triage the ingoing situation and then take action to get your automation initiatives up-and-running—with minimal disruption and maximum throughput.
As you scour the organization for automation candidates, you might instinctively look to offload your most inconvenient processes. Resist that temptation.
Inconvenient processes are typically complex. You want to start easy, not hard: Go with lower-hanging fruit and seek out simple ones to automate.
You should also gravitate toward automating familiar processes, where you can easily identify and resolve issues. This will make your automation more likely to succeed and buys you bandwidth for working on more complex issues.
Think of it this way: Automating bad processes won’t make them any better. It will just make an automated mess.
There’s a lot to be said for open-source code. For one, it lets you leverage the success of others by copying processes that have already been automated. This simultaneously reduces development time and risk.
Look at open-source boards. Seek out processes similar to yours—the closer they are, the better. You want to be able to adapt them to your environment with as little modification as possible.
Low-code/no-code automation has its benefits, too. It allows developers to drag-and-drop functions to build applications, rather than by writing out code by hand. (Of course, low-code solutions allow developers “take the wheel” and write and adapt code if they choose.)
Employee onboarding serves as a good use case for low-code automation. For instance, IT can configure a process that automatically sends an email to a new hire showing them how to set up their PC, phone, voicemail, and so on the moment they join the organization.
While often associated with citizen developers, low-/no-code approaches boost productivity for developers of all skill levels and can be a powerful tool for your IT department. As an example, a citizen developer might use low-code to create three apps in the time it previously took them to create one; meanwhile, in the same timeframe, IT pros with deep expertise might create 12 rather than four.
Robotic process automation (RPA) is a powerful tool that’s rapidly gaining traction in a variety of verticals. It’s ideal for executing traditional, tedious, multi-step processes: think of logging into various disconnected systems, then searching for info, copying, pasting, manipulating in Excel, and so on.
Within that monotonous procedure exists zero room for error—the processes themselves must be essentially air-tight for RPA to work. Similarly, an RPA bot only follows instructions, similar to a basic macro in Excel. Don’t expect RPA to fix a mess or make complex decisions. That’s still your job.
RPA works well for simple, defined processes like procure-to-pay. If an invoice arrives at an Accounts Payable inbox with its data properly and uniformly formatted, then RPA can initiate processing the invoice, handling a number of basic steps without any human intervention whatsoever.
In this instance, you can establish rules pre-approving invoices to be paid if they meet specific conditions. Doing so boosts efficiency while reducing errors.
Despite the hype, automation is no panacea. Provide IT with guardrails on how much data is—and is not—being included in any given automation initiative. Carefully assess the scope prior to launch and institute quality checks before it goes live: You don’t want sensitive information to be “automatically” made public.
Bake in the ability to audit the process and know how to identify potential points of failure. Establish auto-blocks and alerts to notify you when human scrutiny is required. And know how to “hit the brakes” and stop the function before a dire situation arises.
Think of automation like a self-driving car: a human is still at the wheel, alert to the situation and ready to intervene at any moment.
Automation can be deceptively simple-looking. A familiar process may seem like a ripe, easy candidate... until you click “Play” in your dev environment and it quickly goes off the rails. As automation experts, Stratascale can reduce risk and speed deployment of your next project.
Stratascale’s has a center of excellence focused on automation with several different service offerings related to DevOps, cloud and security operations. Please contact us today to learn more.
Jason Hood has almost 30 years of experience running cloud, infrastructure (server, storage, and network), desktop, service desk, security, HPC, and data teams for both Fortune 500 and private equity-funded companies. Most recently, he has focused on the operations side of large data centers, IT modernization, and cloud migrations.