The future of work: five steps to digital transformation in the age of smart machines
Frameworks of artificial intelligence (AI) in action along with case studies and examples are well-described in the book What to do when Machines do Everything: How to get ahead in a World of AI, Algorithms, Bots, and Big Data. The authors - Malcolm Frank, Paul Roehrig, Ben Pring – are from consultancy firm Cognizant’s Centre for the Future of Work. Their earlier book is Code Halos: How the Digital Lives of People, Things, and Organisations are Changing the Rules of Business (see my book review here).
The focus of the book is largely on digital transformation by businesses, and not on work issues like minimum wages, employee reskilling, government regulation, philosophical implications, and ethical considerations. The core of the book is the authors’ business framework with the acronym AHEAD: automate, halo, enhance, abundance, and discovery.
The 12 chapters are spread across 235 pages, and make for a straightforward read. Here are my key takeaways from the book, and my overview in Table 1 (below). See also my reviews of the related books The AI Advantage, Human + Machine, Life 3.0, The Four, The Inevitable, The Industries of the Future, and Machine, Platform, Crowd.
“The rise of AI is the great story of our time,” the authors begin. They define AI as “an area of computer science that focuses on machines that learn.” AI is not just about aping human behaviour, but creating the next generation of intelligence and performance.
The authors explain that the book’s focus is on pragmatic issues and not philosophy or politics. Economic dislocation has always been part of the First (loom), Second (steam), Third (assembly line), and Fourth (smart machines) Industrial Revolution. AI will continue to amaze in the coming years as it will be embedded almost everywhere and in everything.
While reactions to AI tend to be polarised into utopian and dystopian camps, the authors pitch for a more pragmatist view. The age of smart machines is neither a capitalist’s dream nor a worker’s nightmare. Alarmist stories and headlines have “tapped the nerve of insecurity.”
The authors distinguish between three types of AI: narrow (weak, applied, single domain; ‘artificial narrow intelligence’), general (diverse activities; ‘artificial general intelligence’), and super (technical genie). Cloud computing, ubiquitous networking, IoT, and ecosystems of APIs are leading to exponential growth in AI capabilities.
Tech change and impacts proceed through three phases of ‘S-curves’ – innovation burst, stall, and build-out. Rapid expansion followed by maturity are accompanied by bubbles as well as ‘golden ages,’ according to economist Carlotta Perez. AI is now moving from the fringe into the mainstream as new business models and value chains are taking root, the authors explain.
They show that the ‘Republic of Digital’ is now the third largest economy in the world, after the US, and China. The authors present the ‘3 Ms’ framework for digital success: materials (data), machines (AI, IoT, cloud), and models (new business models). New opportunities are emerging for entrepreneurs and corporate innovators to fix problems and frictions.
Digital technology is not just entertaining or convenient, but life altering, the authors explain. AI has moved from “little daily helper” to something more powerful and even disruptive, as shown in games of intellect (chess, Go), driverless vehicles, robotic process automation, radiology, and paralegal work.
Companies need to harness AI for the twin purposes of efficiency and innovation. This can lead to personalised banking, individualised education, effective healthcare, intuitive manufactured goods, efficient supply chains, and cost-effective government services.
For example, GE has reinvented itself as a digital industrial company, thanks to initiatives like the IoT management platform Predix. Education is ripe for transformation; it has been criticised for not doing enough for the top and bottom of the class; the “middle will muddle through” as well. AI along with shuffling of student groups and different offline-online interaction patterns can lead to more personalised and social learning, according to New Classrooms Co-founder Joel Rose.
From factory settings to white collar offices, the transformation of labour is now targeting knowledge activities like codification, recombination, and repurposing of knowledge assets. AI can make knowledge assets richer and lead to interesting and productive uses.
AI’s impact will vary depending on the tasks within jobs, and along different time lines and industries. “Don’t confuse jobs with tasks,” the authors caution. For example, for some jobs, RPA will eliminate only some portions, especially routine and boring tasks. It will make workers more effective, but not necessarily replace them.
“Our prediction is that AI will impact nearly 100 percent of knowledge jobs, while completely eliminating approximately 12 percent of them,” the authors explain. AI will eat existing jobs in a “slowly, slowly, suddenly” manner.
The AI stack
Successful AI implementations focus on the “magic moments,” deliver individualised customer experiences, and suck up data continually from internal and external sources. They connect data sets and empower real-time decision-making at scale via new insights that are actionable and proprietary.
For example, Netflix leverages recommendation engines, the app interface, user data, and AWS infrastructure. It builds on open source best-in-class systems, and can distinguish between what users say about content (ratings) and what they actually watch (viewing behaviour). Amazon has far more API mashups than Walmart, Target or Macy’s, according to the authors.
“Like oil, data needs to be mined, refined and distributed,” the authors explain. Proprietary data, at scale, is the moat of competitive advantage. “Once there’s enough data to draw valuable conclusions, the value of the data set increases exponentially,” they add.
For example, First Data transformed itself from a maker of PoS devices into smart terminals along with management tools and business intelligence insights into customer loyalty. “This is truly becoming an age of knowing it all,” the authors enthuse.
Companies today need to compete at “Google speed” and with “Google price.” Millennials today have “defaulted to digital” and want everything via an app. This is a challenge for incumbents with brands to defend, and large tech and cultural inertia.
“Stability is to be admired, but in the digital economy such considerations can lead to stasis,” the authors caution. Digital transformation helps companies become better, faster and cheaper. “The winning business model will be hybrid – part physical, part digital,” the authors emphasise.
Examples of service transformation include Uber (car as a service), WeWork (space as a service), and GE (industrial uptime as a service).
Companies should not just pay lip service to digital. Creating a standalone app without internal re-architecting is like trying to get to the moon by climbing a tree, the authors joke. Priority should be given to valid objective insights from data rather than just on personality or position of established managers, or even opinions of experts.
Rather than “boil the ocean” of all data, it may be better to begin with specific tasks and quick wins. Different industries are at various stages of digital transformation; though the “melting points” may vary, “digital denial” is pointless since such transformation is inevitable across the board, the authors emphasise.
In a cross-industry survey, they present examples of such process transformation under way in three categories: current, pilot and considering. Examples include wealth management in banks, claims administration in insurance, billing services in healthcare, and merchandising in retail.
II. The AHEAD framework
Automation is one of the first steps in harnessing the power of AI. “RPA is our new loom, our new steam engine,” the authors explain. “Take X and add AI” is the business plan of the next 10,000 Silicon Valley startups, jokes Kevin Kelly, founder of Wired magazine.
There are a number of back-office and middle-office processes ripe for automation, the authors observe. This includes departments like customer service (complaint and help desk management), IT (software asset management), HR (claims management), and finance (compliance).
For example, the Associated Press has been publishing more than 20,000 software-written news stories a year. Emerging startups in automation space are AiCure (clinical data science), Talla (HR) and NextAngles (financial services compliance). Healthcare productivity automation (HPA) provider TriZetto (acquired by Cognizant) automates high-volume routinised work in the healthcare sector.
The authors advise automation of processes that are recurring, routinised, and repetitive. Other candidates are tasks that generate or handle large volumes of data, require low levels of empathy or low demand for human judgment.
Halos of code need to be created by “instrumenting” everything to gather data and deploy decisions, eg. via IoT. Examples abound in industrial machinery (Bosch, Caterpillar, Boeing), athletic apparel (Nike, Under Armour), automobiles (BMW, Tesla, Ford, Toyota) and even insurance (Progressive, Allstate, Travelers).
Every “thing” is now a code generator, the authors enthuse. Code helps “turn water into wine,” as long as the tap of valuable data is always on. At the same time, care should be taken to respect customer privacy and give them a delete button. The role of a digital risk officer becomes important in this regard.
Examples include GE (digital twins), Discovery health insurance (also a wellness company, by promoting healthy lifestyles via wearables and apps), Iora Health (health coaches, outcome management), and New Classrooms (design thinking in classroom activities).
Enhancement of human performance is possible via smart tools, data insights and automation. From the stone age to the digital age, human performance has been continually amplified by tools. Tools have lifted us “from the savannah to the Sea of Tranquility” on the moon, the authors explain.
Higher velocity of decision-making can lead to greater user gratification. Using technology to automate some customer-facing tasks opens up opportunities for greater creativity by customer service staff, eg. in Zappos, Apple, Pret A Manger. Applying the human touch, bringing a sense of fun, displaying empathy during tough customer moments, and showing generosity of spirit are tasks that are best done by humans, acting along with smart machines.
ImagineCare’s predictive services enhance the competencies of doctors. McGraw-Hill’s ALEKS (Assessment and Learning in Knowledge Spaces) establishes a unique pathway and cadence for each student, with assessments in real-time and not just at the end of the semester. Playing against AlphaGo even led champion Lee Sedol to come up with new types of moves.
In terms of physical capabilities, exoskeletons have been built by Panasonic (for lifting heavy weights), Ekso Bionics (to help paraplegics walk), and Sarcos (for defense personnel). For office workers, enhancement is offered by Signac (detecting unlawful trading), and Avicenna (improving radiology diagnosis).
Abundance or democratisation of goods has been made possible by successive waves of innovation over the centuries: clothing (loom), travel (steam engines), and manufactured goods (assembly lines). What were once regarded as luxuries became affordable and ubiquitous. The authors predict that similar trends will appear in financial services, education and entertainment, thanks to smart machines.
“Obsess about the startup community,” the authors emphasise. Market maps of startup activity help business leaders understand the transformations in fintech, healthtech, insurtech and govtech. Even though some of the initial offerings may look inferior, flawed or even irrelevant, there is a lot to learn from the successes and failures of startups, the authors advise.
Makerspaces are other sources of useful insights, blending engineering, design, creativity, and entrepreneurship. “Innovation is inherently messy, but it needn’t be chaotic,” the authors explain.
There is both an art and a science to innovation. Useful planning approaches are McKinsey’s three horizons (current, short-term, long-term), or even the LVMH model (60 independent subsidiaries, which may even compete with one another).
Provocative exercises suggested include asking your sharpest employees how to put your own company of business, or envisioning how the business would be restructured so that your current premium offering is sold for free or only at 10 percent of the price.
A strong focus on metrics also helps (“digital Taylorism”). For example, sensors, software, process re-design and analytics have helped Narayana Health Care make heart surgery more affordable without loss of quality.
Discovery via R&D, process analytics and portfolios are key for innovation management in the digital economy. Such initiatives should include top-down and ground-up approaches. Just as Edwin Budding’s invention of the lawn-mower in 1827 led to a boom in cricket, rugby and tennis via grass courts, so also smart machines will open up vast new unimagined possibilities.
“Innovation at the velocity and scale of AI” is the need of the hour, according to the authors. Humans learn at linear rates, while machines are learning at exponential rates, according to futurist Ray Kurzweil.
Companies need to blend incremental innovation (digital Kaizen) along with moonshots in their portfolio. They need to think like VCs, with hits paying for the misses. They will also need to learn to be comfortable with failure as a source of learning; after all, the movie and music industries have had their share of failures as well. Even mighty companies like Microsoft and Apple have had some failed products (Vista, Lisa), while others like RIM and Nokia have faced a reversal of fortune.
For example, the University of Kentucky conducts surveys on students for a predictive scoring system which helps improve retention rates. Toyota is investing in driverless cars of the future, for which “AI is the central technology” – as well as investing in traditional models.
Humans will have to continue to ask questions, be curious, imagine and build, all using the new machines, the authors explain. Looking back at how much the world has changed also helps get a sense of perspective, eg. the Internet Archive Wayback Machine. Humans have a hard time envisioning or extrapolating into the future, and believe they are at the summit of progress rather than at a high camp on an infinitely high mountain.
Leaders must be energised by the unwritten future rather than just trying to hang onto the glories of the past, the authors urge. AI is the ultimate game changer, and inaction will lead to irrelevance, the authors sign off.
Also read: The Top 10 Books of 2018 for Entrepreneurs