Digitalisation vs. Automation: it is about capabilities

A pre-digital organisation focuses on efficient processes implementation; a digital organisation focuses on capabilities.

When working out a transformation strategy these days the difference between automation and digitalisation always comes up. The usual argument is that the organisation has already automated all processes that can be reasonably automated and for the remaining processes there simply is no automated alternative. This premise is based on the view that digitalisation is about processes, when in fact it is about capabilities.

All process optimisation methods before digitalisation are indeed about processes: a process is a sequential, procedural arrangement of tasks that plays out over time while a capability is an inherent attribute of an organisation.

Methods for process optimisation

A manual process evolves through electrification to an industrialised process; an industrialised process evolves through mechanographication to an information process. An information process becomes an autonomous process through automation. And an autonomous process ceases to be a process and becomes a capability through digitalisation.

Digitalisation is the last transformation stage of a manual process which leads to the process dissolving and being replaced by one or more organisational capabilities.

Electrification adds speed and efficiency to a manual process and results in an industrial process

Process	Industrial process
Evolves from	Manual process
Method	Electrification
What	In an industrial process work is performed by machines which are operated by people, information is consumed and produced by people and decisions are made by people.
Example	I call my manager on the phone and ask for vacation. I then submit a vacation request form to HR. HR makes a photocopy of the form and files it.

Electrification is “the process of powering by electricity” and, for the better part, that’s it. Electrification speeds up a process and makes it more efficient, enabling miniaturisation and allowing for mass application of that process.

manual process + electrification → industrial process

To be precise, any sufficient power source when applied to a manual process through machinery can lead to an industrialised process. However, the process evolution can not progress to digitalisation without the unique benefits of electricity.

Switching a process from a manual/chemical/thermic energy source to electricity comes with many benefits:

• Electrified processes are efficient as there is little energy waste
• Electricity consumption and production are spatially separated
• High speed of electrified processes
• Electric applications are flexible
• Because of electricity’s high energy density, electric devices are compact and powerful

Thanks to electricity numerous manufacturing and communication processes were electrified during the last century, but also many new applications became possible that were earlier not feasible due to “classic” power source inadequacies, greatly aiding manufacturing, communications and data storage.

Electrification doesn’t do more than adding efficiency and speed; it doesn’t include any automation or processing intelligence per se and does not replace human actors in a process.

Mechanographication adds information to an industrial process and creates an information process

Process	Information process
Evolves from	Industrial process
Method	Mechanigraphication
What	In an information process work is performed by machines which are operated by people, information is consumed and produced by machines and people and decisions are made by people.
Example	I send HR a vacation request email, CC-ing my manager. My manager checks with the team calendar and replies to the email. HR checks with the HR system.

Mechanographication is a kunstwort which I’m fairly certain doesn’t exist in English, maybe exists in French but most positively exists in Greek, the cool words’ club house, were it means “recording of data with machines” (mechane, μηχανή → machine, graphe, γραφή → writing).

Actually, one can record data with a mechanic typewriter or a phonograph (or anything ending in -graph), but there isn’t much fun in that without the benefits of electrification. Mechanographication leverages the efficiency and speed of electrification to record and categorise large amounts of data with little effort and in short time, building the basis for all information processes. Following the proud wordsmithing tradition of phono/tele/photo-graphy, I’ll broaden the definition of mechanographication to include all recording of data with machines, including telemetry and sensory input eg. from digital cameras.

industrialised process + mechanographication → information process

When mechanographication is applied to an industrialised process, information enters the equation and the process becomes an “information process”.

Automation adds decision taking to an information process and creates an autonomous process

Process	Autonomous process
Evolves from	Information process
Method	Automation
What	In an autonomous process work is performed by machines, information is consumed and produced by machines and decisions are made by machines.
Example	I submit a vaction request through the HR system. The HR system validates my remaining vacation days. My manager receives a notification. My manager checks team availability and approves the vacation. I receive the approval.

Automation takes an information process and re-implements it with machines, eliminating manual tasks. An automated process thus doesn’t depend on manual input to run and thus is autonomous. An autonomous process is driven by machines that perform tasks, but also take decisions based on the information available to them. It is the decision taking part that sets autonomous tasks apart from information processes, which also include machines and information. The states and transitions remain the same, the information passed around remains the same and decision points are formalised as conditions which machines use as input for decisions. As machines generally work faster, more efficient and precise than humans, an autonomous process usually outperforms a manual one.

information process + automation → autonomous process

However there are preconditions. Automation requires an orchestration of the technical components participating in the process which, in turn, assumes a degree of interoperability. This is similar to people who collaborate in a process: they need to be instructed and trained first to the process, even if they are experts in their field, before they can collaborate with others. The effort spent on process automation thus fundamentally depends on how easily the involved components can be made to collaborate with each other; the more effort we spend on making components talk to each other, the harder automation becomes.

Autonomous processes encounter failure just as manual processes do and it is often that error handling has not been thought through. For all our weaknesses, humans are reasonably OK when dealing with unexpected situations; machines are, yet, not so good at that. A central decision in process automation is what to do about errors: either harden a process to the point that errors become rare enough that they don’t pose a threat or increase robustness so that errors can be dealt with easily in either an automated way or through cheap, manual intervention.

Automation heavily makes use of mechanographication to get data into the process and electrification in order to execute it efficiently. Automation always mechanises an existing, manual process, innovating in execution. However automation does not invent new processes.

Digitalisation turns an autonomous process into a capability

Process	Capability
Evolves from	Autonomous process
Method	Digitalisation
What	A capability is not a chain of tasks but an organisational attribute.
Example	I visit the HR system in order to requestion vacation. The HR system, taking into account team availability and remaining vacation days, shows me available dates. I then commit my vacation request through the HR system. The HR system approves my vacation request and notifies my manager.

Digitalisation leverages automation to reinvent business processes that would not be possible without automation. In that aspect, digitalisation is to automation what electrification is to manual processes.

autonomous process + digitalisation → capability

Good candidates for digitalisation are processes where decisions can be taken based on the available data and only that and where the digitalisation benefit outweighs, at scale, penalties of wrong decisions. For example, a digitalised procurement process for less than 500€ could be executed automatically because the process has all information available, is fairly cheap to execute through the ERP and there is a 14 day free-of-charge return policy for the procured goods in case the need for cancelling the order should arise. Thus, a procurement process becomes an availability capability.

Just as electrification requires electricity, digitalisation relies on four cornerstones of modern information processing: IoT, big data, AI and cloud.

• The “internet of things” collects global data from the real world in real time and feeds information and actions back into it, creating the basis for data-hungry digital systems.
• “Big data” summarises the technical capabilities necessary to process large amounts of data in a meaningful way – much more data than a person or a team of people could process in a short amount of time and provides the foundation for computationally intensive data processing algorithms.
• “Artificial intelligence” builds on big data and establishes cognitive capabilities on par with – and often exceeding those of – human experts.
• The “cloud” is the technical service model with which digital capabilities evolve cheaply, flexibly and speedily so that IT services keep pace with business development. The cloud provides the reliability and performance necessary for 24/7 digital capabilities.

Epilogue

Digitalisation changes an organisation through the use of technology. The interaction with technology becomes transparent, predictable and repetitive tasks are carried out by machines as people concentrate now on novel and creative tasks.

Contrary to automation, where a process is first designed and then automated, digital processes are designed with an organisation’s digital capabilities in mind. Digital processes cannot be executed meaningfully in a manual way as they enable capabilities which require cheap, fast and reliable task execution. The process fades out as capabilities move into focus. A pre-digital organisation focuses on how to optimise and execute processes; a digital organisation focuses on the capabilities it needs in order to achieve its goals.