In my previous article, I explored what intelligent automation is, why it's receiving such attention in business and industry, and what the four most important success factors are.
Now, I want to turn to the practicalities of deciding what type of automation is best suited for the job − there is no point in buying a Rolls-Royce if a Datsun bakkie is really what's needed!
The point is that intelligent automation is not necessary for all automation projects. For example, many automation projects are required to operate within fixed parameters that will never or only infrequently change.
Artificial intelligence (AI) or machine learning (ML) would be unsuitable for such projects because they are designed to respond to changing parameters all the time. In fact, when the parameters are specified, too much intelligence would tend to be disadvantageous.
Functions like case management, content services or good old robotic process automation continue to have a place in any automation strategy. In short, it's all about selecting the right tool for the job.
The pitfall is that there is so much hype in the marketplace about AI tools and the amazing things they can achieve, that technology-inclined executives may opt for too much intelligence.
Conversely, when used selectively, AI-driven intelligent automation tools can open up new opportunities where automation needs to be underpinned by complex reasoning and constant adaptation.
Making the right automation decisions
Alan Pelz-Sharpe and Dan Lucarini, leading authorities on intelligent automation, argue that the ideal start to an effective decision-making process is understanding what the current or ‘as is’ processes are, and what they should look like (the ‘to be’ process).
Automation is widely misunderstood as substituting humans with machines.
Pelz-Sharpe and Lucarini identify seven steps to making the right automation decision:
- Analyse the organisation's existing business processes, steps and tasks, and evaluate them for potential improvement. Three questions need to be asked: What is really happening at an IT / data level? What tasks and subtasks do employees (or customers) follow to get things done? What is the high-level conceptual ‘journey map’ of customers and employees?
- Deploy process mining tools to automate the discovery of what really happens to data at the system level.
- Create a journey map that marries the system-level picture with the human-level one but be aware it may still not provide the full story.
- Design the ‘to be’ process. This is the easiest and most fun part of the whole exercise − it's where the team's creativity and ambition are given full rein.
- Identify all the people, processes and technology that will be involved, even tangentially.
- Get advice. This is all about choosing the right tools and, more importantly, the right partner in the form of a consultant or value-added reseller. Intelligent automation is a specialised area, and getting the right advice is critical, both for scoping the project and choosing the right tools.
- Make the business case, focusing on value. If planning to use intelligent automation backed up by AI and ML, it's important not to be distracted by the shiny new technology and all it promises. The focus should be on the quality of the data and desired outcomes.
These steps are not sequential, nor are they discrete. For example, depending on the level of in-house automation skills, it might be wise to get a partner on board early in the process to help with the initial design of the ‘to be’ state.
Don't ignore the human dimension
Automation is widely misunderstood as substituting humans with machines. In fact, nearly all automation projects are intended to augment human capabilities rather than replace them; less frequently, humans may be used to augment the work of automation technology. But the point is that automation is about bringing humans and technology together.
One should therefore consciously think about any automation project in terms of improving the business, making it more competitive and strategic, and not about reducing headcount.
While technology can do many things − such as processing high volumes of data quickly and accurately − better than humans, humans can do many things better than any algorithm.
Therefore, say Pelz-Sharpe and Lucarini, the key to success lies in establishing what the right balance between human and machine is at the earliest possible stage. The later in the process this happens, the more expense incurred in implementing changes and even scrapping technology.
One of the biggest learning curves when e-commerce came into vogue was that trying to replicate the real-world experience online was ultimately self-defeating. The real winners were those who were prepared to go further and imagine a new kind of retail experience that leveraged the new possibilities opened up by the internet.
The same type of dynamic has informed subsequent technology developments, such as the shift to mobile. Organisations contemplating a shift to automation, especially intelligent automation, should take this lesson to heart.
A truly successful automation won't just recreate what already exists − though that might be a first step − it will look to new opportunities that a partnership between humans and technology could open up.
Share