Leslie Willcocks

Professor Emeritus

London School of Economics and Political Science

“Which businesses are more likely to gain a lead with automation and digital?”

We have been studying this for the last 9 years, and a clear picture has emerged. The top attribute is that they have an executive team that believes that digital technologies can be transformational and of strategic importance to its business as whole. Actually, the evidence is that these technologies, managed well, do have these characteristics. The more digitised businesses are, the better the financial results they get in their sectors. In fact, getting ahead and being a leader in these technologies is likely to give the organisation a competitive advantage that can become irreversible over a 5–7-year period.

Such an executive team will fund and resource the technology as a long-term strategic investment to build a digital business platform that delivers necessary internal changes much faster, can support new product/service and can seize business opportunities as they arise.

In our book 'Becoming Strategic with Robotic Process Automation' we detailed all the strong attributes of a digital leader. Getting there, we suggest, involves:

1. 1. Think, then act strategically—the business imperatives dictate your technology investment.
2. 2. Start right—pilot but build always with the end technology platform, architecture and infrastructure in mind. Build the organisational and technical capabilities for delivering and deploying the technology.
3. 3. Institutionalise fast—we had transformation managers who refused to start until they had all the stakeholders on board and the governance structures in place.
4. 4. Remember that if it’s strategic, it involves major organisational change, and change management really is the key to the door here, and not just an add-on. At the heart of it is a major cultural change that will take several years to happen.
5. 5. Innovate continuously. For some examples, have a look at Amazon; DBS Bank Singapore; Schneider Electric; Google; TenCent; LexisNexis. Their digital platforms allow for faster, easier and more extensive innovation both internally and in external relationships and business positioning.

“What Is intelligent automation, in the scheme of things?“

Over time, especially as more advanced technologies emerge, it became obvious that stand-alone RPA was essential, but offered unnecessary limitations. Intelligent automation consists of more powerful software suites that combine RPA with more advanced technologies—typically business analytics, OCR, intelligent character recognition, natural language processing algorithm and machine learning based software—that can automate or augment tasks that do not have clearly defined rules. We have also called such technology ‘cognitive automation’ but, as we mentioned above, really do not like to call such software ‘Artificial Intelligence’ because we believe the AI label aggrandises what these tools do. With these technologies, inference-based algorithms process data to produce probabilistic outcomes. A variety of tools are in the realm of intelligent automation, such as tools that analyse data based on supervised machine learning, unsupervised machine learning, and deep learning algorithms. While some of the algorithms have been around for decades, only recent advances have provided the computational power needed to execute them on big data. The input data is often unstructured—such as free form text, either written or spoken. For example, at Deakin University, Melbourne we found intelligent automation being used to answer natural language student inquiries. The input data can also be highly structured, such as the pixels in an image. Google’s Machine Learning Kit, IPsoft’s Amelia, IBM’s Watson suite and Expert Systems’ Cogito have been older examples of IA tools.

So, to summarise, where were organisations with this in 2023? We are seeing organisations increasingly combine RPA with what they call ‘AI’—computer systems that seek to simulate and outperform human intelligence, for example analysis, logic, memory, processing—machine learning using structured semi- and unstructured data, computer vision and natural language processing tools, and process discovery and mining, whereby data analysis is used in order to detail and improve business processes. The potential applications are vast, and exciting—witness for example the immense interest in Chat GPT from early 2023—but therein lie also at least three problems: The first is choice. Which processes and applications do we go with? The second is management. How do we productively build the capability to develop, manage and deploy these technologies? The third is how do we control the power and possible downsides, and unanticipated consequences of what we create? It is worth signalling a fourth issue, which we touched on in an earlier question, because it is proving a real stumbling block in a number of large organisations we have been recently researching.

The challenge is data. As much as 80 percent of an organisation’s data is ‘dark’, meaning that the data is un-locatable, untapped, or untagged. Enterprises first have to create new data and clean up dirty data that is missing, duplicate, incorrect, inconsistent or outdated. Enterprise adopters of IA tools also struggle with ‘difficult data’, which we define as accurate and valid data that is hard for a machine to read, like a fuzzy image, unexpected data types or sophisticated natural language text.

Now intelligent automation/’AI’ applications generally require large data training sets, and thereafter are set up to deal with massive amounts of variable data. Processing power and memory race to keep up. If a great deal of data is not fit for purpose, then this bad data will create misleading algorithms and results. The idea that very big samples solve the problem—what is called ‘Big Data’—as used in ChatGPT, for example, is quite a naïve view of the statistics involved. It is not really possible to correct for bad data. And, as we said before, the dirty secret of Big Data is that most data is dirty.