With the rapid growth of automation, has Process Management had its zenith?
Process Management has become the lens through which industry is understood by the majority who own and work in our organisations. It underpins advice given by consultants, forms organisations around its workflows, drives cost reductions and introduces efficiency. It has allowed industrialised production, playing a major role in globalisation and poverty reduction for developing countries, alongside contributing to some of the largest problems that we have ever faced as a species.
Origins of Process Engineering
The mass adoption of process organised work was formed around the industrial revolution. This division of labour, went hand in hand with the introduction of machines. At the turn of the 20th century Frederick Winslow Taylor championed the concepts we understand today: function specialisation, defined methods, decision governance, etc. The effort to describe human work in these terms enabled mechanisation. Defining standard processes allowed specification of machinery, standard interfaces allowed interaction between machines and people.
For the majority of the 20th century this scientific management of work remained in the domain of industrial engineering. Towards the end of the century the growth of computing and the popularity Six Sigma, TQM and Lean saw this reach expanding into other functions. Computers presented an industrial-revolution scale opportunity to realise efficiency, not just on the work-floor but in management and service functions. Process re-engineering offered not only predictable efficiency in services, but the opportunity to define this work in a manner that computers could consume. The new millennium saw an explosion in process re-engineering as a tool for reducing costs through efficiency, automation and outsourcing.
So what’s changing? Automation!
… In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it. The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters. In the Deserts of the West, still today, there are Tattered Ruins of that Map, inhabited by Animals and Beggars; in all the Land there is no other Relic of the Disciplines of Geography.
purportedly from Suárez Miranda, Travels of Prudent Men, Book Four, Ch. XLV, Lérida, 1658
There must be a limit after which process re-engineering ceases to be effective. To be successful a company must innovate; they must explore. The ability to alter course, strategy, methods or product gives the opportunity to differentiate from competitors. We intuitively accept an “edge” to activities that can realistically be defined in the scientific manner: where the effort to define process doesn’t produce the returns. We perceive on one side of the line activities that can be predictably recorded, on the other those that require skills such as judgement, creativity, etc: the “human” skills. When we look into the former group, we see the activities that could or have been defined as processes, and thereafter those defined processes that are automated.
We can imagine it like the shoreline. The beach are those activities that are theoretically manageable through “scientific method” with the sea of automation lapping at the edge, making gradual inroads as the tide comes in. We imagine there are some folk on the beach who need to be careful less they get their feet wet, but that the majority of us are sat in the beach cafe, happily located on the promenade. We tell ourselves our activities are too complex to be defined predictably, let alone automated. We believe the water line can never reach us.
We have internalised the notion that automation is limited by what can be described in a scientific manner, because until this point that is all our machines have been able to consume. We instinctively draw a distinction between dumb computers with their binary approach and smart humans trading in ambiguity. This is a wholly incorrect assumption.
DLaaS (Deep Learning as a Service)
The increase in cloud-based, distributed power that can be utilised by commercial organisations for AI computation, aligned with the increasing consensus on standard languages and frameworks to define and train AIs, gives industry the tools it needs to commercially utilise Artificial Intelligence. The significance of this cannot be underestimated. Companies now have the resources available to train AIs in activities deep into our metaphorical promenade. These resources are available on an industrial scale and becoming cheaper. It is well within the reach of any listed company to run a section of their IT dedicated to the development of AI based solutions. This was unthinkable just a few years ago.
If this sounds like more of the same automation, think again. The difference between old-school rules-based automation and deep learning is stark. The AIs learn based on data, comparing their guesses with data to better approximate future outcomes. The absence of a need to specify rules to transform input to output – essentially the essence of process engineering – allows automation to step off the beach and firmly plant a foot on our “safe” promenade. The human and machine limitations that saw the growth of process engineering – finite capacity of the former, inflexibility of the latter – no longer exist with Deep Learning that is fully scalable in activities containing ambiguity.
Automation consequences for the workforce
First it means that a current limitation on the adoption of automation – the speed at which we can actually transform to a process oriented way of working – is no longer a hurdle. Functions that should be in scope to automate, but haven’t been due to the challenge and complexity of mapping the activities, can be attacked from a different angle. If they are theoretically good fodder for process engineering except for the challenge of scientific definition, they likely also have the data available to train AIs. That means the step of defining as rules can be skipped, we allow the AI to infer the connections in its internal organisation.
Second, it means that the those who interface the previously automated functions – preserving a role through defining their skills as a required human interface point – are also good targets for AI. Functions that require repetitive classification, decision making, diagnosis or judgement calls – where the output is given with a confidence interval rather than as an absoute certainty – are excellent opportunities.
At first glance this sounds like a very limited set of jobs, but in reality it covers a massive swathe of human activity in the workplace. The vast majority of roles are somewhat repetitive, interspersed with grey-area judgement calls. This presents challenges to organisations: how do we set ourselves apart from the competition? What will be our USP? And crucially: what happens to a workforce that has spent the last century being taught to deliver in a scientific manner, only to find that the computer has removed the need to do so? Have we unwittingly placed much of our workforce on the beach, watching as the tsunami approaches?
In the next blog we’ll look at how Change Managers, used to framing transformation in context of process management, need to shift their thinking to help organisations and individuals address the new challenges.