The economic and operational uncertainty caused by the COVID-19 pandemic has forced many companies to focus on quick automation wins. It seems that most of those successes have come from task automation (RPA) without a focus on the big picture and the business processes they are part of.
This will likely create long-term process and technical debt for organizations. To alleviate this scenario, a new approach is proposed, which is Hyperautomation.
The initial challenges are the following:
- How to scale beyond RPA and achieve such hyperautomation
- The Automation Alignment Matrix for Hyperautomation
- Hyperautomation roadmap
Definition of hyperautomation
Starting at the beginning, business-driven hyperautomation refers to an approach where organizations quickly identify, examine, and automate as many approved technology and business processes through a disciplined approach.
Business-driven hyperautomation involves the orchestrated use of multiple technologies, tools, or platforms.
A model to deal with the hyperautomation process would fit the following scheme.
Step 1:
The first step is to define business goals. What will be the result? What needs to change? Is growth being pursued and revenues need to increase? Increased profitability? Are costs reduced? Do I have to apply a compliance policy?
For example, for Wells Fargo, compliance, regulation and legal have been one of its main drivers. They require moving towards risk reduction and ensuring that government regulations are followed.
When you start with the commercial objective, you drive what ends up being the digital ambition. This ensures focus on key issues. It's always about goals and technology as support and help.
Step 2:
Once the commercial objectives have been set, a global vision of the needs of the existing programs and processes is established. Most of the improvements and changes will be automation, as many jobs have substantial elements that can be automated.
According to WorkMarket, 53 percent of employees believe they could save up to 20 hours a month by automating tasks (WorkMarket 2020 In (Sight) Report).
Restricting yourself to automation is a partial vision that needs to be expanded and practices such as Design Thinking or Game Theory can be used to take the company through what-if scenarios. It may be that you derive a vision of completely tearing down what currently exists and automating the entire process from scratch, or iteratively improving what you have by partially automating the process.
This conclusion is quite common since most of the processes are designed for other technological, commercial moments, or do not contemplate the new digital habits of the market and customers.
Step 3:
With business goals and vision, you must go through the process of thoroughly documenting the work to be done. A previous analysis of the processes helps to align the results of the processes with the commercial objectives.
Fortunately, this job is no longer as difficult as it used to be. Automated process discovery and data mining tools have evolved remarkably for this.
Still, technology is a means, not the end. Subject matter experts are needed to work with process analysts to take their models to the extreme.
Automated process discovery and mining tools have inherent limitations that still require human intervention and analysis to give you the information you need to design your processes in ways that align with your digital ambitions.
In particular, process mining tools attached to RPA tools almost always end up drifting towards RPA. If you have a hammer, everything looks like a nail.
It should also be considered that it is time to perform scenario simulations to see what results appear as changes are made to the processes. What if instead of manually sorting invoices, a document ingestion engine does it? What if there are bots taking over some routine work?
Step 4:
Once an extensive and ambitious definition of the work to be done is available, an analysis of technological solutions is carried out to determine which is the most appropriate for the type of expected result.
This analysis, which is more visual if presented in matrix mode, is necessary as they often had difficulty determining when to use which of the many available automation tools.
The DigitalOps toolkit is a comprehensive foundation of the components that should appear in the matrix. The alignment of automation is a very important step that ensures that the problem mentioned in Point 3 of ending up drifting everything to RPA-only solutions is avoided.
This analysis is a vital step in ensuring that you are using the right automation tool at the right time and place in your processes. It could be argued that it belongs to the Process Analysis step (Step 2), but it is sufficiently relevant and complex, taking into account the infrastructures available in each organization, to have its own “step”.
Steps 5 and 6
Once the preliminary tasks have been carried out and it has materialized, at least in regards to RPA and the implementation of bots, first prototypes or the creation of minimum viable products to determine settings more effectively and efficiently.
This step is already well established and is a good practice throughout the industry. The same goes for release and production, at least for initial deployments. Where there are challenges is being able to scale beyond prototypes, pilots, and limited production deployments.
Step 7:
All the exposed steps, carried out in an orderly manner, are key. Nevertheless, monitoring and measurement are key in order to establish the scalability plan and grow the efforts towards Hyperautomation.
Monitoring is beneficial to help arrive at the result. However, it is of little use if you do not know where you are starting from, where you are at this moment.
It is a map that cannot give instructions or make future predictions if it does not know the context.
In the automation process, none of this is possible without an effective Data Analytics model.
Trading information is not a good indicator on its own. The real situation and its monitoring must also be able to measure the performance of the solution, change management, incidents, costs, etc. With that information you can predict what will happen.
Essentially, all of this data should inform and forecast whether the results of your process will help or hinder you in achieving the business goals defined in Step 1. This is what really matters. If you're not getting the right process results, figure out how to make the appropriate change to achieve the desired result. The rest of the performance metrics are interesting but not decisive.
Step 8:
As the first successful results begin to appear, a governance and scaling plan is required.
The challenge is that, at this point, there are many disciplines required for the variety of tools available in DigitalOps, as well as less technical but no less critical disciplines such as process analysis or reengineering.
Additionally, as scaling occurs, new automation efforts will also be incorporated.
The methodologies and the COE (Computer-Optimized Equalization, practices for balance between technological elements) must take this entire scenario into consideration.
Attempting to scale automation processes without a government that establishes methodologies and strikes a technological balance results in duplication of effort, silos, and disorder. In other words, you may be automating, but the approach will be highly inefficient across the business and therefore have a high opportunity cost versus competitors.
Step 9:
Continuous iteration, review and improvement is vital, and an integral part of driving business forward.
This often raises the question of when to stop spending money on process improvement. The answer has to be clear: when the process is considered so good that you no longer need to worry about the competition, then you can stop improving. This generates a next question that everyone in the organization must ask themselves is if they have reached a level far above the rest or, more ambitiously, if that level of optimization can ever really be considered.
On the other hand, basing this decision strictly on competition is an oversimplification. There is also the consideration of business objectives, resources that are not infinite, customer satisfaction, trends, ... all these factors force a selection of certain processes in which it is necessary to commit to continuous improvement.
Hyperautomation is described as an infinite loop.
Conclusions
Early successes in automation are often basic RPA implementations. This generates a lot of general excitement, but then it slows down on how to scale and grow these partial successes all the way to Hyper-Automation.
The most important drivers of ultimate success, once you get past that initial euphoric phase, are the left two-thirds of the infinite loop schematic shown in the schematic in this post.
At that moment, the little experience in Hyperautomation is manifested. Steps one through four, and seven through nine, are where the ordered, planned, iterative part of the infinite loop comes into play. These steps are where the heavy lifting is. They are the disciplinary part of Hyperautomation.
However, they also have the most significant reward for their efforts. The best way to distance yourself from the competition is to link process results to business goals and adjust based on those results. To achieve this, a holistic approach to automation must be taken, and not just focused on one third of the endless loop of hyper-automation.
In short, hyperautomation is a business-driven, iterative, and disciplined approach that accelerates automation efforts and results.
Hyperautomation cannot be achieved with RPA alone. The DigitalOps suite is key to being able to have the necessary tools to scale. While automation efforts are often successful in initial pilot implementations and in small departments, organizations need to scale beyond initial successes. Getting to hyperautomation requires a more rigorous approach to defining your business goals up front and letting them drive the digital decisions you make as you build your automation program.
