TechUpdate: what is hyperautomation?

Hyperautomation is a new automation philosophy or strategy in which organizations automate as many business processes as possible end-to-end by combining multiple technologies, such as RPA (Robotic Process Automation), AI, machine learning, process mining, and low-code/no-code tools. Hyperautomation differs from conventional automation in several respects.

In the first place, scale and scope are central to hyperautomation. It is not about a task or a process, but about automating entire workflows, including the more complex, non-routine steps that previously required human insight. It can therefore involve end-to-end processes. A second difference from classical automation is the explicit, combined deployment of various technologies. A third difference is – as also stated Gartner Research – that artificial intelligence is applied in hyperautomation. The prefix 'hyper' therefore refers to the scale and intensity compared to ordinary automation.

How it works

Hyperautomation is not about automating easy/repetitive work, or the technical contradiction between deterministic software and AI. It is about a much more fundamental question: you automate as much as you want to automate. Often, organizations focus solely on cutting out the easiest and most repetitive tasks, or they go too far and want to blindly leave everything to bots. The true essence is consciously retaining the interactions that deliver real value to the customer: the moments where the 'human touch' makes the difference. You can automate all the rest of the work.

A hyperautomation platform consists of various solutions and is capable of determining the best route and steps based on the process, situation (information about the context), and other variables. If a death is reported to a life insurance company, you handle this differently than when an address change is received by an e-commerce platform. Automated solutions are available for this triage as well.

In customer contact – but also in other areas – managers are strongly focused on gaining an advantage using AI applications. Many organizations start deploying point solutions without a strategy. That is good for gaining initial knowledge and experience. However, those who want to fully capitalize on the opportunities must give AI a place in their IT and automation strategy.

Why is it relevant

Hyperautomation is the next step in the pursuit of efficiency, if only because of the scarcity of human staff. In customer contact, therefore, the question is relevant as to where the deployment of human staff is desired and/or necessary. However, hyperautomation can also improve the customer experience, for example by eliminating waiting time or speeding up processes.

Four ingredients are indispensable for handling customer contact: knowledge of the organization, knowledge of the customer (the context), communication, and executing actions. The latter two tasks can be performed by humans or (partially) automated. In doing so, a clear distinction is made between the use of generative AI (for a flexible, natural conversation) and deterministic technology (for a fixed, reliable output). Organizations would do well to choose the right technique for different parts of a process. GenAI excels for the conversation (conducting a dialogue, drafting emails, or formulating responses). However, for implementing changes or actions in underlying systems, where a 100% predictable and traceable outcome is required, tight, rule-based workflows are the only logical choice.

Just pay attention

1. The choices you make regarding hyperautomation – which technologies you use to automate which processes – are determined, among other things, by the need for control over the process. Non-deterministic technology (such as GenAI), for example, is not good. auditable, unless additional aids and restrictive solutions are deployed.

2. As mentioned, a second point of attention is the role of the employee. Is it necessary (for example, due to laws and regulations), desirable (due to marketing strategy or quality), or redundant? Choices in this area can vary per (sub)process.

3. A third point of attention is that costs can be unpredictable when deploying generative AI. These costs can vary due to the workload per process, but also due to the number of workloads (more customer requests, more processes being processed) and changes in supplier billing models. When using deterministic solutions, the costs are known in advance and are often lower.