“Going digital” has long been touted as a silver bullet for delivering better customer experiences and streamlining processes. Automation has become the go-to approach for solving immediate pain points, mainly in the form of tactically deploying one-off robotic process automation (RPA) initiatives to make our jobs easier and/or more efficient. Some of these implementations have been very successful.

But the reality is that many organizations are still struggling with “automating from tactical”. This is not to say that there aren’t good reasons for this approach. Some organizations, for example, may not have the technical competency or the requisite tools and skills to source and implement automation without the support of a consultant. Or they may simply not have the staffing to handle the continuous deployment of new technology. It’s certainly better than doing nothing at all.

However, the reasons why this approach may not be a practical choice in a given situation may have less to do with the technology itself and more to do with the context in which it is being deployed. “Automating from tactical” – or at least adopting an approach that doesn’t necessarily promote a “one-off” model of implementation – needs to be placed in its rightful context. The consequences of not doing so could have unintended consequences when it comes time to reap the benefits of being an early adopter when technology advances can no longer be used to justify current levels of automation.

So, let’s take some time to unpack this concept and explore how each of these four approaches might actually play out for your organization.


1. Tactic 1: Automate All Manual Work With RPA

The goal here is to achieve a “one-off” level of automation that would reduce a small or specific set of manual tasks that otherwise don’t get automated because they are not required at the moment. A good example of this is replacing a long-running, manual process with an automated process – say, the manual provisioning of transportable computers for use in an office. This “tactical” approach typically leads to highly manual and inefficient processes. For example, once this process has been automated, there will be no manual provisions, but the automation itself requires a manual trigger that requires a user to toggle between systems and approve each transaction manually.

Another problem with this approach is that it adds more automation to the IT environment than is required. This will quickly create bottlenecks and reduce visibility into production-level governance and control of the application landscape, and it may even render the changes difficult to maintain. This is because production changes may impact other areas where those changes were not intended. From an operations perspective, this borders on chaos as you’ll have multiple instances of master data (i.e. mobile devices, desktop, file servers) and applications that are managed through manual processes interacting with systems that are managed through different automated processes.

The concept of the “tactical” approach to automation has been further fueled by the advent of AT&T’s Project Maven and the related “Automation From Tactical” white paper. Both initiatives essentially attempted to bring RPA-based automation to a tactical level of application management and distribution, and they were quite ambitious in scope and operationally sophisticated in nature. While both initiatives fared well on their own terms, they quickly drew criticism from many quarters for attempting to automate everything at once.

2. Tactic 2: Automate All Software Development With RPA

In this approach, automation is used as a means of accelerating development workflows and increasing speed to market. In particular, organizations can automate more granular tasks such as build, test, and deployment as opposed to the more overarching workflows used in the first approach. The advantage here is that automation is applied on a task-level basis and will help teams optimize their development effort through standardization, configurationization, and specialization of tasks that are repetitive or that require a higher level of expertise. This approach has worked very well for large organizations that have taken a “vertical integration” approach to automate various aspects of their software development life cycle (SDLC).

The problem with this approach is that it can create pockets of automation where automation was not necessary. For example, let’s say that standardization has been achieved by automating all common build, test and deployment tasks for a given application. In theory, the next step in the evolution of this process would be to automate all custom versions of these activities, which would require a much higher level of automation skills and technology. In practice, however, large organizations have been overly cautious when making this shift from vertical to horizontal integration in terms of accelerating development workflows and increasing speed to market.

This approach also has the potential to quickly convert programming staff from being a bottleneck to a bottleneck, as they are the ones who have the knowledge and skills necessary to implement automation. In other words, it can create a “knowledge vacuum”, as those who need to manage these processes will be required by those responsible for the development process.

3. Tactic 3: Automate Everything With RPA

The goal here is to automate as much as possible with the goal of eliminating manual processes that do not require high-level skills or specialized tools and therefore aren’t worth automating at all. The idea is that IT departments can get all their automation under one roof, resulting in more visibility into the application landscape and increased efficiency. This approach sits on top of all previous approaches, by expanding the number of tasks that are automated through an “all-in” approach.

An error with this approach is that it requires a focus on automation without any consideration for the context in which automation is to be deployed. For example, many organizations that have transitioned to an “all-in” approach are still carrying over manual processes for reasons that are not readily apparent to the operations team. This can be particularly problematic for organizations that are not adept at maintaining automation, as they will quickly start to lose visibility into the effects of their automation efforts on production environments. In most cases, manual processes can be minimized or eliminated altogether through a combination of RPA and AI/ML technologies, but the biggest challenge will be finding a way to maintain those manual processes until they can be replaced by automated processes.

4. Tactic 4: Automate All Software Development With RPA And AI/ML

Here, automation is used as a means of accelerating the development of application-level tasks and driving innovation. For example, it can be used to accelerate software development for new version-level capabilities, allowing more opportunities for teams to push forward with more creativity and specialization. For example, it can be used to accelerate redeployment of applications at scale; where manual testing typically requires specific individuals with knowledge of the application landscape who are often on short-term contracts. By applying automation to testing, automation can eliminate some or all testing entirely. There are many other examples that fall into this category as well (e.g. “automate all integration, automation and testing”).

The advantage of this approach is that it can provide transparency into the application landscape that’s not possible otherwise (i.e. the ability to see every version of an application deployed to production), resulting in increased speed to market and increased customer satisfaction. The big problem is that automation requires a level of knowledge and skill that cannot be easily replaced by automation, which has resulted in many organizations taking on excessive automation at scale (resulting in automation “pockets” where unnecessary labor exists). To make matters worse, organizations are finding ways to get out of “automation at scale”, which increases the reliance on AI/ML solutions, which are equally ill-suited for this task.

In conclusion, the good news is that there are tactics to follow to automate every aspect of software development, depending on your organization’s situation. The bad news is that even if you follow all the tactics above, there will likely be automation “pockets” where labor exists that are unnecessary for your particular organization. At the end of the day, it’s all about where you see the greatest value in automating your development workflows and reducing the labor required for building and maintaining applications.

Interested in reading more about how automation is used in businesses today? Automation Becomes Essential in the Post-Pandemic World discusses the importance of automation in businesses today.

AutomationIf you would like TSVMap™ to assist your business with assessing your essential systems and applying the TSVMap methodology to ERP SystemsMRP SystemsCyber SecurityIT StructureWeb ApplicationsBusiness Operations, and Automation, please contact us at 864-991-5656 or info@tsvmap.com.