When it comes to digital transformation (opens in new tab), there's a lot of attention on cloud computing (opens in new tab) and the implementation of AI (opens in new tab) and machine learning (opens in new tab). But as Ky Nichol points out, automation is also a key consideration:
There is no dispute that automation (opens in new tab) remains an instrumental driver of technology and business decisions, data insight, continuous delivery, and business benefit. Organizations are therefore designing strategies to drive efficiencies and unlock the potential of their workforces beyond the expectations of repetitive, manual processes much better suited to machines. Implementing effective automation is challenging at the best of times, not least when embedding it across business processes and operations, spreading its impact from the traditional view of responsibility held by technical teams.
Even if we aren’t familiar with the Disney façade of dancing brooms (for those unaware of this reference to Disney’s ‘Fantasia’, The Sorcerer’s Apprentice depicts this really well!) and a spiraling loss of control from a ‘simple’ process, it is all too common a fear that the implementation of automation could spiral out of our grasp. This is an area that many organizations trip up on. For effective automation, you need to not only ensure that you get the right tools for the right job, but that you optimize them in the right way across the organization, too.
We can (very) loosely liken this concept to the process of baking a cake. When we’re talking about automation, I’m not talking about bringing in the best blender, secret ingredient nor star baker. I’m talking about the ultimate recipe. The hub, the complete set of instructions across all human and machine activities that, with the right orchestration, result in a cake anybody would be proud of. This formula impacts all pieces of technology and provides a solid grounding to that end baked result - yes, the decisions on blenders, whisks and hidden ingredients are integral - but we need a meta layer to factor in conditional branches (i.e., a blender stops working, or is no longer needed - we need a whisk!) and to oversee the whole process for this to run effectively. Then, as the process takes hold, those branches could extend to better blenders, more advanced ingredients and wider components added over time.
What does this mean in terms of the wider sense of automation implementation? How can organizations bring in automation in a manageable, measurable way without the risk of disruption and chaos? How can we embed that across the organization in a manageable way and what does that realistically look like?
This is where the ‘automate-first’ culture comes in. To enable teams to deliver, programs across the organization need to be architected with humans and machine automation running simultaneously. In today’s highly competitive, dynamic and unpredictable world, businesses need to be able to act and respond quickly and have watertight processes and repeatable measures in place to do so, and in that context, it is inevitable that multiple automation tools will be in place across a business. The role of the ‘automate-first’ culture is to join up these tools and build processes and flows of work across them.
There are three fundamental factors to an ‘automate-first’ approach that we need to consider here, and I’ll explore in more detail:
Crucially, automation should originate from within the team it impacts, and embedded into the jobs of those responsible, while identifying the roles of humans and machine as complementary. My experience of environments such as the space industry highlighted that, in many senses, we could tip our approach to automation on its head, and start thinking about automation-first, and then how human capability is intertwined. In spaceflight, automation is not only a given, but a necessity, and humans are then involved for the more nuanced, less machine-led responsibilities which allow both parties to operate at their optimum level. From a culture perspective that means we need to identify those processes and tasks that can be automated, thereby removing humans from repetition and tasks better suited to machines.
By shifting responsibility to more creative and thought-led processes, the human workforce can focus on what they do best, and on a level augmented by automation. This also reduces that risk of loss of control and shifts operations to a movement more akin to an orchestra, with every player and instrument working together.
The other reason that humans do not always fare well with repetitive, manual processes, is that human error is all too common. That’s why automation is so beneficial, whether as a full replacement or a useful aide. That said, the nuances of human knowledge and experience are specific to particular individuals within teams and organizations, which can often be lost in translation, historic interactions, or through staff attrition. That’s why documentation and knowledge sharing is such a crucial mitigator and guardian of insight. The introduction of automation in combination with human insight in this context enables teams to prepare for failure and build and maintain that knowledge-sharing structure.
The orchestrator of orchestrators
Our learnings from working with leading industry organizations is that what’s needed is a better architecture to set out governance, control, visibility and toolset autonomy, otherwise known as a unified automation strategy. This would look like an enterprise layer infiltrating the full organization and tying the threads of automation centrally. We have identified this as the orchestrator of orchestrators. The orchestrator, or the conductor, in some sense, pulls these pieces of tech together in agreed sets of activity to achieve objectives that you can use as a foundation to optimize automation in a given process.
As technology evolves, organizations will undoubtedly have a collection of existing tools and skillsets forming pieces to the puzzle of a hybrid automation model fit for purpose across the business. Central to the ‘automate-first’ culture must lie a way to join up these tools and build flows of work across them that permeate the organization. This central system, the ‘orchestrator’, might not automate everything, but acts as the mission control to reach out and trigger the specialist systems across an organization that automates particular activities. This allows you to build value quickly, without having to discard any existing toolkits.
How to take care with a force multiplier.
Automation is a force multiplier. In that sense, we need to be really careful about how it is introduced to ensure that we maintain control and there are no unintended dancing broomsticks, in the context of our earlier Fantasia analogy. We often talk about operating ‘clutch control’, from an operational standpoint especially, to ensure that capabilities across the organization, from marketing and sales to finance and operations are coordinated. Though we don’t want to slow the pace of change and improvement, it’s important to operate on the same level and with the right visibility to maintain control, and operate in lockstep to translate that change into great business outcomes.
Similarly, it’s very tempting to pick the ‘easy bit’ when it comes to automation, or to focus on one tool and ‘bend it out of shape’ beyond its original purpose. There’s nothing wrong with doing this as a start, but when production starts to scale, the danger of a bottleneck or a tight spot becomes very real. You can’t pick and choose for too long in isolation, and it’s important to have a full view of your environment at the outset, whilst maintaining focus on the question of ‘why are we automating?’ clear.
This is, again, where the holistic approach comes in. At Cutover, when we talk about clutch control, that clutch links the world of 400+ releases per day in development to the wider organization. In the world of DevOps (opens in new tab), it is all very well investing in a Formula 1, best-in-class engine, but when the supporting IT infrastructure (opens in new tab) isn’t there (cue 1970’s mini chassis for the sake of argument) to interpret that power, the balance is skewed, the power is not linked to great steering and brakes and that additional energy dissipated. The balance is therefore key, and needs to come at a ‘low tax’ on the pace of DevOps whilst still fostering control, governance and visibility across the business. Technology change is all about business outcomes, aligning the development engine with the steering and brakes of the organizational framework, to navigate to effective operational value. It’s on that basis that automation needs to be instilled across the organization in a purposeful way, avoiding that loss of control and knock-on loss of efficiency, if the risks aren’t factored in properly.
With so much to factor into an effective automation strategy, the three points outlined certainly provide a foundational grasp of the issue, but still sit at the tip of the iceberg. At Cutover, we have been spending a lot of time with leading organizations and our own experienced practitioners to drill down into the core of automation and this is an exciting time for our industry in terms of its capabilities and potential. It’s clear that there are several threads tying the success factors of automation implementation together, but I do think that culture and orchestration (opens in new tab) lies at the heart of it. By redefining an organization’s approach to automation, whether it be via roles and responsibilities, the pace of change, legacy factors, or unifying processes, we are able to build a more progressive, visible change across business models that can truly drive growth and add value.
- Ky Nichol is CEO of Cutover (opens in new tab).
- We've featured the best cloud storage (opens in new tab).