What To Do When Machines Do Everything

Book review

What To Do When Machines Do Everything
by Frank, Roehrig and Pring
published by Cambridge University Press

Reading time: 4 mins

Malcolm Frank, Paul Roehrig and Ben Pring are senior consultants with US technology strategists, Cognizant.  Their 2017 title, What To Do When Machines Do Everything, presents an upbeat case for a future economy in which artificial intelligence (AI) has been widely adopted.

This assessment foresees a medium term net increase of 1% in the total number of full-time jobs, in contrast with the gloomier forecasts of many contemporary thinkers, including Martin Ford, whose 2015 best-seller, Rise of the Robots, predicted the onset of mass unemployment.

What To Do… has won many plaudits, including book prize nominations, so it’s worth examining the authors’ contribution to the AI debate.  At just 200 pages, it is brief but also compelling.

Authors: Ben Pring, Paul Roehrig and Malcolm Frank

Technological revolutions have played a pivotal role in economic history.  Humanity has progressed through the use of tools made from stone, bronze and iron.  Our latest tool, they argue, is data, and its historical impact is destined to be equally significant.

The book’s fundamental purpose, it seems, is to guide and motivate business leaders through the hurdles of transforming their organisations into data science-led corporations.  This data science focus (sometimes called ‘big data’) will be the key to their survival.

The opportunities are substantial, because AI has the capability to reveal structures and patterns in vast, diverse data sets which an army of experienced analysts using conventional tools would be incapable of identifying.  Its capacity to create new products and even new business models is already in evidence.  As Flavio Villanustre of LexisNexis explains, “By adding a data set to another data set, you can potentially make a completely new thing.”

AI enables greater personalisation of goods and services, so traditional standardised product ranges and conventional production methods will need bottom-up transformations.  Employees with a flair for innovation must therefore be identified and given the freedom to take risks.

One effective self-improvement technique is the so-called ‘kill the company’ workshop, where senior managers are invited to build a narrative explaining a hypothetical future in which their business has failed.  New mindsets, policies and re-structuring plans should emerge.

In order to thrive in the new economy, companies must firstly generate and capture as much data as possible about their own operations, their assets, products, markets and customers.  Inevitably, there will be too much data for any one person or team of people to analyse, but this is where AI steps in.

The authors claim that in the next decade “the breakout companies will be those that become masters in consistently turning this abundant data into actionable, and proprietary, insight.”

Hence, leaders will need to trust data as the basis for making decisions.  The days of inspired hunches and HiPPO* interventions will be banished forever, although the authors warn that the “brass wall” of middle management will try to resist AI’s increasing influence.

* A HiPPO is a 'highest paid person's opinion', often experienced when a senior manager in attendance hijacks a group decision, especially during meetings.

So, What To Do… makes the case for prompt and decisive action by the leaders of established corporations, otherwise their businesses might simply be ‘killed’.

At this point, let’s introduce a note of caution and consider the interests of the authors, who are all senior figures at Cognizant.  Any book advising established corporations to grasp the opportunity to exploit AI now, before it’s too late, could be interpreted as a sales pitch for more advisory work:  “Where does this leave the 100-year-old companies, or even the 40-year-old companies?...Actually, in a very good place…if they move quickly.”

If the authors had concluded that the old corporate giants are already doomed to fail (and beyond help) in the face of their fully digitalised, upstart competitors, we would surely have applauded their honesty.  But that isn’t their conclusion.  Hence, readers of the book will need to decide for themselves whether this note of caution (hopefully not cynicism) is justified.

It’s also worth spending a moment to consider the implications of this thesis for government decision-making and public services.

If Cognizant are correct, AI’s ability to generate wealth through innovation, operational efficiencies and quality improvements across all sectors should result in improved tax revenues.  What’s more, there will be a reduction in levels of unemployment.  All of this should enable government to reduce debt, redistribute wealth and invest in infrastructure.  It seems almost too good to be true, although the history of technological revolutions demonstrates that this outcome is plausible.

So, who is most likely to benefit from reading this compact and accessible book?  There is little doubt that business leaders, middle managers and IT professionals need to take cognizance (no pun intended) of what Cognizant have to say.  Indeed, anyone interested in the implementation of AI within organisations will have much to think about long before the final page is turned.

On balance, What To Do When Machines Do Everything is a recommended read and a pleasant antidote to the dystopian fears of many commentators.  However, if there’s one incontestable flaw with this book it’s surely the title.  If machines are expected to create more jobs than they destroy, often simply augmenting the work of skilled humans, then how can it be said that they will Do Everything?

What To Do When Machines Do Everything is published by Cambridge University Press and available from all major retailers.

Writer: PJ Moar of Moar Partnerships
Email: p.moar@moar.com
Twitter: @MoarPart