IT automation is paradoxically old and new.
It’s “old” in the sense that various forms of scripting, business process automation, and other technologies have been around for decades. It’s “new” – or newer, at least – in terms of containerization, Kubernetes, infrastructure as code, security automation, robotic process automation (RPA), and the vast landscape of cloud-native tooling.
No matter what, it’s tough to find an IT shop taking a “just say no” approach to automation in 2022. On the contrary, there’s a trove of numbers that all seem to say: IT automation is virtually everywhere. That’s evident in software pipelines, hybrid cloud infrastructure, security operations centers, you name it.
IT automation is no longer just for IT, either, as departments throughout an organization look to leverage tools such as RPA – often enabled by IT, but not entirely dependent on it – as part of their own broader transformations.
While automation has grown immensely of late, it’s far from all grown up. This is an area defined by continuous growth and optimization. Moreover, even “automate everything” teams are still optimizing and improving.
[ Related read: Automation: 5 issues for IT teams to watch in 2022. ]
Nearly three-quarters of respondents (74 percent) in In Red Hat’s 2021 State of Kubernetes Security report said they have adopted DevSecOps. In the same survey, 25 percent of organizations said their DevSecOps implementation is in an advanced stage that integrates and automates security throughout their software pipeline – indicating lots of exciting progress and plenty of room for continued growth.
5 pieces of automation advice for IT leaders
Whether your team is in the earliest phases of an automation initiative or an “automate everything” organization or somewhere in between, we’ve culled five pieces of expert advice to help.
1. Begin incremental, become strategic
A key piece of automation’s appeal is that it can very much be done incrementally: A Bash script here, an automated web app vulnerability scan there – at whatever pace your capabilities allow for.
“This is automation through the lens of traditional system admins and even site reliability engineers in many cases,” Red Hat technology evangelist Gordon Haff told us recently. “Do something manually more than once and automate it so that you don’t have to ever again.”
You don’t need a big bang launch; automation done right will always be a work in progress anyway, so why treat it otherwise?
But Haff and other experts note that incremental approaches to automation should not preclude planning. Incremental makes it manageable; strategy will make automation successful.
Leon Godwin, principal cloud evangelist at Sungard AS, advises creating and using a digital strategy document that defines key aspects of your automation approaches, especially in any scenario where automation is one of the major levers of an organization-wide digital transformation.
That document should define “the experience they want for their internal and external stakeholders and how that differs from the experience they have today,” Godwin says. “They should also identify the barriers inhibiting them from achieving this outcome and what technologies could help overcome these barriers.”
Your strategy should also identify the expertise, technologies, and partners that will be needed to achieve your automation and digital transformation goals, according to Godwin.
Your strategy should be agile and flexible enough to pivot when things aren’t going well, Godwin says – which makes the incremental approach a good fit, too.
2. Treat ROI as cumulative and compounding
Successful organizations obsess over ROI for a reason: If you spend a lot of money on hiring, technology, training, or any other dimension of your business, you want to be able to see (and hopefully celebrate) the results.
But a lot of ROI measurement and discussion happens in narrow or even artificial parameters, such as on a quarterly or annual basis, or on a per-project (or similar) basis.
Take the long view with automation ROI, especially if your implementation is incremental.
“Worry less about the ROI per automation you are trying to tackle,” says Ton Roelandse, managing director at Trexin Consulting.
Roelandse adds that when you break automation down at the task/process level – as you would in many RPA workflows, for example – many might flunk a simple ROI test of “time to build the automation” versus “time currently spent per week per resource.”
But when you bundle all of your manual tasks together across different processes, it’s another story – that’s when you’ll see, as Roelandse says, “the automation light.”
The same basic principle applies widely to various forms of IT automation, whether in a CI/CD pipeline, a cybersecurity strategy, hybrid cloud management, and elsewhere. “Ultimately it is the compounding value of automation that will drive your value,” Roelandse says.
3. Develop automation and data skills everywhere
As noted above, automation is no longer solely about IT. It’s everywhere (or soon will be in) in the modern organization.
As the amount of work in an organization conducted by machines and code grows, there will be a corresponding need for people who can act appropriately on the outputs of automation, machine learning, and AI, according to Amaresh Tripathy, global analytics business leader at Genpact. Most companies aren’t there yet.
“For most organizations, small teams with specialized training are responsible for making these decisions today,” Tripathy says. “In the digital age, however, this model is unsustainable – all employees must have access to the tools and skills they need to unlock the power of technology.”
[ Learn how leaders are embracing enterprise-wide IT automation: Taking the lead on IT Automation ]
The advent of things like low-code or no-code tools that make it easier for automation to spread and even start outside of IT doesn’t diminish this need – it accelerates it, since long-term automation success will depend on a wide variety of people who understand the inputs and outputs of increasingly automated workflows.
Genpact itself had work to do in order to practice what it preaches. “For example, we recently launched an initiative to increase data literacy across our organization and shift employees away from transactional projects and into insight-generating roles,” Tripathy says. “In essence, we aim to equip about 100,000 employees with the data and analytics tools, techniques, and skills they need to drive value for our customers.”
4. Treat data management as more than a technical matter
In highly automated environments, good data management and governance is far more than a technical issue – it can rise to the level of moral obligation, not to mention a matter of regulatory compliance, brand reputation, and more.
“An often-overlooked challenge in automation is the efficient, secure, and ethical management of data – especially for enterprises in highly regulated industries like financial services,” Tripathy says. “Prioritizing security, compliance, and removing potential data biases is crucial.”
This is another area that can longer be the sole purview of a relatively small team. The wider and deeper your automation strategy runs, the more important it becomes to have people across a wide range of roles who understand the importance and implications of the data they work with.
“This effort requires a concerted alignment across the enterprise as part of an overarching data-driven transformation strategy,” Tripathy says. “It means creating clear guidelines for data management and upskilling employees for increased data literacy with ethical usage and regulatory compliance top of mind.”
5. Understand how data flows through your organization
You can distill a lot of automation goals – and automation challenges – into two categories: Process and data.
With process (which some people might break down into the smaller unit of “tasks,” or roll up into a larger workflow consisting of multiple processes), one of the common requirements is to ensure that you have a complete understanding of and visibility into said process, both now and later. Otherwise, you run the risk of helping a bad process run faster and more frequently – or not run at all.
With data, one of the common requirements is – wait for it – to ensure that you have a complete understanding of and visibility into said data, now and later.
“Go into any digital transformation exercise asking the question about how that data is going to flow around the enterprise,” says Craig Stewart, chief data officer of SnapLogic. “Each vendor is normally only concerned with the direct integrations from their app to others, not the wider picture of all the dependent inter-related applications.”
Just as data literacy will be crucial to long-term automation ROI, so will data visibility. In fact, the latter is a prerequisite for the former.
“Someone has to have that higher level view, and understand where the potential bottlenecks in this automated system are,” Stewart says. “Make sure there is one place where this can be seen, and one place for the organization to go for resolution of any new challenge."
[Where is your team's digital transformation work stalling? Get the eBook: What's slowing down your Digital Transformation? 8 questions to ask.]
