Contact us
We are taking you to another website now.


Taking a Machine First Approach to Digital Transformation


Many companies believe they can compete in the Business 4.0TM world merely by digitizing their existing products and services, and the key business processes that create demand and supply for them. That would be a mistake. Instead, leaders should take a step back and consider how they can reshape their entire business around digital technology - thus reinventing their business using what we refer to as a machine first approach to digital transformation.

By "machine first," we do not mean replacing every possible human worker with artificial intelligence and other technologies that automate labor. Instead, the machine first approach asks leaders to determine how to use three key digital technologies—cloud computing, artificial intelligence and analytics—in every aspect of their business: in the products and services they offer, how they develop their offerings, how they create supply and demand, how they support their customers, and more. By combining an agile approach to developing and continually improving these products and processes, companies can become digital powerhouses with new data-fueled business offerings that delight customers.

It is how a growing number of companies have become leaders in their sectors—digital natives such as Amazon in ecommerce and cloud services, Netflix in streaming movies, Uber in taxi service, and Peloton in fitness equipment. But it’s also the way that companies founded before the rise of the Web are now overtaking competitors: CVS in pharmacies and health clinics; ABN AMRO in European financial services; Cummins in the truck engine market; and many others in industries that have rapidly been digitizing.

To take a machine first approach to revamping their business, executives must look for ways to eliminate unnecessary manual work of both the production and knowledge types. But leaders also must then free up employees to take on new roles and learn new skills. At the same time, they must reimagine and reposition their IT function, making its focus business-first, and its core capabilities, ones we describe as ubiquitous, polymorphic, and secure.

In this article, we explain in more detail why companies should take a machine first approach to digital transformation, how to do it, and the capabilities they will need (especially their IT function) to make it real.


Reshaping the Business

In the Business 4.0 world, many products and services are becoming a purely digital phenomena, accessed through a digital device. Think streaming movies to your smartphone or laptop computer, or finding homes to buy and obtaining a mortgage through your tablet computer, or making a stock trade, or purchasing a ticket to a play or concert. They’re increasingly becoming digital offerings secured through digital means.

In this transformation, technologies such as automation, cloud computing, and the Internet of Things (IoT) devices enable companies to make unprecedented (and swift) changes to their business models. They also allow companies to make major improvements in key business processes—marketing and selling, product development, production and distribution, order fulfillment, talent management, and more.

As companies increasingly compete in a Business 4.0 world, they need to take a machine first approach to rethinking and redefining every aspect of their organization. This would include reexamining the following:

  • Business models - What they sell, to whom they sell it, and how they generate revenue from the transaction. For example, $20 billion engine manufacturer Cummins found a new revenue stream by digitally tracking the post-sale performance of its engines and providing remote monitoring that alerts trucking customers when their vehicles need preventative maintenance. Launched in 2014 as Connected Diagnostics, the service tracks more than 120,000 Cummins engines today in North America—more than 20 times the number just three years ago.
  • Core products and services — These increasingly will shift from analog alone to digital, or a combination of the two. For example, privately held stationary bike manufacturer Peloton has reached $700 million in annual revenue and a $4 billion valuation not only by selling its $1,995 bikes for home use but also by offering a subscription (for $20 a month) to live-streamed exercise classes. These subscriptions have played a part in helping Peloton achieve more than 100% annual revenue growth since its launch in 2012.3 For this stationary bike manufacturer, the product is no longer just the bike, but also the fitness exercises that come with it.
  • Marketing and sales — Advances in analytics enable companies to mine big data and unearth insights about the products and services that customers need and want, and how they use and react to them. Companies can use these findings to develop tailored promotions that are more likely to achieve sales goals. One major telecom company was able to reduce customer churn significantly by tracking customer account data and creating customized offerings to retain their most at-risk customers.
  • Production and distribution processes — Multiple Business 4.0 technologies are converging to improve and reshape the supply chain. Improvements in analytics make it easier than ever to track every item at every link in the chain, helping to spot fakes and prevent theft. 3D printing enables companies to move production all the way downstream, even to a customer’s home or business. Synchronization among manufacturing, distribution and order fulfillment makes it feasible for companies to produce customized products for individual consumers. In pharmaceuticals, for example, advances in genomics and precision medicine increasingly enable the development of medicines tailored to an individual’s specific genotype, which have huge potential to be more therapeutically effective with fewer undesirable side effects.
  • Customer Service — AI technologies are helping contact center reps and field service technicians deliver higher-quality service. However, at the same time, smart, self-correcting, connected IoT products may reduce the need for human interventions. In other words, products ranging from cars to barbecue grills will increasingly fix themselves. Sophisticated, data-driven chatbots are handling many routine customer inquiries, thus helping companies reduce the costs of call centers by obviating the need for human customer support. Where human customer service reps are needed to deal with more complex situations, they can process calls better and faster when working hand-in-hand with virtual agents that have already captured all the critical information on the customer's issues.
  • R&D — Data analytics can guide product development teams by showing them which innovations are likely to make the biggest splash in the marketplace. Remote analysis of real-world product performance can give marketing managers valuable insights on ways to improve quality and customer satisfaction. Sales force teams can use analytics to get location-specific visibility into how well supply and demand are matching up in their territory.
  • Talent management — AI also can assist with everything from hiring and development to engagement monitoring to find, motivate, and retain the best employees. AI will increasingly be used to determine which roles in a company are far more important than other roles, and the effectiveness with which people in those roles operate.
  • Finance and accounting — Analytics can help companies accelerate cash flow by automating accounts receivable and accounts payable while analyzing expenses to find ways to reduce costs.

Many companies will be tempted to use technology to make incremental improvements to their existing business processes. But a machine first delivery model calls on companies to rethink each element of their business.

Using the powerful AI, automation, cloud, analytics, and IoT tools at their disposal, companies have a golden opportunity to rethink their business model, which customers they should target, what products and services they should develop for those customers, as well as how they should create demand and manage supply.

Many companies will be tempted to use technology to make incremental improvements to existing business processes. But a machine first delivery model requires rethinking each element of their business.


How to Implement a Machine First Delivery ModelTM

The first step is to consider new business models that would allow the company to compete within a digital ecosystem that includes not only major players like Google and Facebook, but also hundreds of smaller competitors and startups.

Next, companies should look into introducing new products and services. In all cases, the focus should be treating technology as a means to achieving desired outcomes, not as an end itself.

For example, the retail industry is ultimately about delivering goods to customers in their homes. Among other benefits (such as promoting its Prime Service) Amazon saw an opportunity to use its 2017 Whole Foods acquisition to enhance the value of its Alexa virtual assistant by allowing users to order groceries via Alexa, while establishing distribution centers for the groceries that those users order.

In the automotive industry, manufacturers such as Ford, BMW and Toyota are responding to the growth of the sharing economy and companies such as Uber and Lyft. They are looking for ways to expand beyond manufacturing and supplying cars toward reinventing themselves as mobility companies that get people from A to B by using IoT systems to monitor automobiles and make software fixes/downloads automatically (as Tesla, among others, have begun doing). And to keep customers happy, automakers have started dispatching technicians to repair cars at customers’ homes and offices, rather than making customers disrupt their day by driving to a service center and leaving their car for repairs.

This essentially amounts to a digital redefinition of an industry. It will have huge implications for companies whose products can be digitized 100% (think media, entertainment, and financial services). In the real estate sector, mobile app provider Zillow has moved from providing home price estimates to generating leads for real estate agents. That’s easy for Zillow to do given that its site has 151 million visitors every month.4 More recently, the $1 billion (revenue) company has been partnering with banks to provide mortgages to its users. And in August, it entered the mortgage origination business itself through an acquisition.5

It’s useful to think about Zillow’s digital journey—from data provider to mortgage underwriter. It entered the digital marketplace in 2006 as an online database of homes in the U.S. It produced a highly popular mobile app that let home buyers to preview homes for sale as they walk or drive about, street by street. Now, with more than 100 million app users,6 it is providing more value to its home buying shoppers: helping them get through the onerous mortgage loan process. That is a huge opportunity for Zillow, especially given that the share of U.S. home mortgage money loaned by the three biggest U.S. banks fell from 50% of all new mortgage lending in 2011 to 21% by September 2016. The 10 biggest mortgage lenders by volume are now non-banks (e.g., Quicken Loans). Only two were in the top 10 in 2011.7

Financial services firms around the world are encountering uncommon, and often digitally unfettered competitors. And many established banks are fighting back successfully.

After determining a new business model (as Zillow did for its business), the next step is to design and build the business processes necessary to support those new products and services. Here the focus always should be on delivering a superior customer experience, for example through AI-enabled and digitally sophisticated chatbots that can handle issues by knowing customer histories and respond accordingly in a human manner.

Of course, it’s important to put security at the heart of any new digital offerings or business process. The machine first delivery model relies on collecting and analyzing massive amounts of customer data. That data needs to be protected, and it’s much better to integrate security into system architecture from the start rather than trying to bolt it on after the fact.

One benefit of cloud computing is that patch management and security updates can be delivered remotely and automatically to distributed systems. This can help companies maintain regulatory compliance while protecting both software and hardware from bad actors.

As companies develop these new systems, models, products, services, and processes they will need to adopt agile, automated systems to minimize disruptions to existing core businesses.

But the next step will be to empower and reskill the IT function. That will enable IT professionals to play an even bigger role in the transformation of their company. We’ll explore that in more detail next.

Where We Tested the Machine first Approach First: On Ourselves

For years, TCS leaders have believed there is no better way to test a new consulting approach than to test it on ourselves. Toward that end, we launched several initiatives in the last two years to refine and improve our machine first delivery approach to digital transformation. Those initiatives have already had significant impacts in two areas: talent development and energy usage.

Let’s look at talent development. Since we are in the business of helping large, global companies transform themselves with digital and other technologies, it would be very easy for our employees’ skillsets to become outdated. Since 2010, a number of technologies have been turning out to be crucial to many companies: big data and analytics tools, digital sensors and other Internet of Things technologies, cloud computing, and artificial intelligence, among them. And yet the supply of highly skilled people in those arenas remains relatively small, as it does for any new or fast-evolving technology.

In addition to finding that talent, TCS management also believes highly in retraining current employees. It’s a philosophy that has helped us more than double our revenue since 2011, to $19 billion, and double our headcount, to nearly 400,000 today.

We have trained 260,000 of those employees to improve their skills in these emerging digital areas through digital learning. We took a machine first approach beginning in 2015 to internal training and development. Prior to that, internal training was largely through traditional classroom lectures at our facilities worldwide. We realized that this would be the slow, highly expensive and often not highly effective approach to learning new consulting and technology skills.

As a result, we have redesigned our training and development processes from scratch. We took our classroom training content and converted it into online modules. We now have an immense digital learning platform with more than online 350 courses and 139 guided tutorial sessions. (From that platform, we issue 470 certifications to our people.) Over the last nine months alone, some 23,000 associates in 23 countries have developed new competencies in key digital areas such as analytics, Internet of Things, cloud, agile, and machine learning.

Online training—delivered just when an employee needs it (and not when a classroom session is offered), and at a convenient place (wherever they are sitting)—has proven to be hugely successful to TCS and its employees.

The second place in which we have applied a machine first design approach is energy management. Our company has a huge electricity bill—in the millions of dollars. (We have 120 large buildings across India from which we deliver many of our services. In all, we have operations in 46 countries, and nearly 150 delivery centers.) Our Indian facilities collectively house about 275,000 TCS people—about 10% of the population of Chicago—and they work in aggregate space that is five times the capacity of the world’s tallest building: the 160-story Burj Khalifa in Dubai.

In total, TCS’ more than 120 facilities in India were consuming 500 MWh units of electricity annually. By using machine first principles and technologies such as IoT, cloud, and analytics, we instituted energy-saving practices such as automatic shutdowns of light and heat in unoccupied spaces. This year’s total electricity bill will remain the same as last year’s even though we have increased our floor space 15%.

We have many other machine first initiatives at TCS, including in human resources (especially in using social media to recruit new employees across the world), finance and our internal IT function (we have 50 data centers across India). We expect the returns on them to be just as substantial, and in some cases, even more.

The Machine First Delivery Model: Four Keys for IT Management

One of the biggest transitions for a company to make in a machine first digital transformation will be in the IT function. Specifically, the IT group will need to make four fundamental changes in the way it operates (if it hasn’t already done so):

Making data ubiquitous. The first aspect of making data ubiquitous in a company is identifying all the data it needs to run a superior business. Much of this data is in the hands of other parties in its ecosystem: suppliers, partners, customers, and other external parties. Walmart’s ecosystem includes suppliers (e.g., Procter & Gamble) and distribution partners that are key elements in making products available in the right quantities and at the right times in Walmart stores. Once they seamlessly combine such ecosystem data with internal data, companies that design their business with a machine first approach do not hoard or isolate that data. Instead, they give real-time data to employees at all levels, when and where they need it. As one CIO of a large pharmaceutical company has said, the way to avoid "the finger-pointing between business and IT on data is by making data available to everyone."8 But data alone is not sufficient. The IT function needs to provide analytics tools, as well as the insights about the company (and customers, suppliers, etc.) that those tools generate, throughout the company. For example, having deep insights about exactly why customers buy from a company is crucial to that company’s marketers, sales force, customer service agents, R&D, and finance (if the pricing group sits there), among others.

Embracing polymorphic analytics and other uses of digital data. Following the Greek meaning of the term (having multiple forms), polymorphic in this context is about tailoring analytics applications to the needs of managers and other workers, function by function. Each C-suite executive in a company has his or her own key metrics. The IT function must tailor data analytics tools to each of their needs. For example, analytic tools for CFOs should bring unprecedented insights on cash flow, margins, and risks. A chief operating officer would need analytics tools tuned to such metrics as the customer experience, customer satisfaction, and on-time delivery of orders. Then, those executives must be trained on how to use the tools. If they aren’t, the insights are likely to go untapped. Here’s an example: At the Mall of America in Minnesota, mapping Wi-Fi data allows mall retailers to identify customer movement through the facility and optimize their offerings to attract them, as well as use weather data to predict mall traffic. But such data is only valuable if the mall’s business users (including store managers) become “power users,” as the Mall’s IT director says.

Maintaining a focus on customer needs. As machine first companies embrace advanced technologies, they must guard against getting so wrapped up in the wizardry of technology that they lose track of their core mission: to satisfy internal and external customer needs and exceed their expectations. In all cases, a machine first delivery model should also be business-first; technology should be deployed first in the service of the business, not to make things easier for the IT function (although a lot of simplification needs to be done there too).

Embedding security in every system. Because the machine first delivery model is based on the proper use of data, companies that adopt the approach must take extra care to make sure their data is protected from both external and internal threats.



As machine first companies embrace advanced technologies, they must guard against getting so wrapped up in the wizardry of technology that they lose track of their core mission: to satisfy customers.

Machine First Won't Mean Machine-Only

Employees have understandable concerns that AI, machine learning, robotics, and other technologies that enable automation will eliminate their jobs. However, history suggests that new technologies create more jobs than they destroy. In the 1800s, agriculture accounted for almost 75% of jobs in America. The mechanization of farming associated with the industrial revolution did cause a 25% decline in farm-related jobs by 1890, but it did not lead to widespread unemployment, because new jobs opened up in other fields—including jobs associated with the design, manufacturing, sales, and repair of industrial farm equipment.

By 2030, for example, McKinsey Global Institute predicts that automation could displace some 400 million workers worldwide, but that other catalysts including rising demand for health care from aging populations and new investments in infrastructure and energy should still generate demand for up to 890 million new full-time equivalent positions.10

It is true that some companies generate more revenue today than they did decades ago despite employing many fewer workers. At the same time, even leading digital-first companies like Amazon are investing heavily in retaining and adding employees.11 In the first quarter of 2018, Amazon employed more than half a million people, up more than 50% from the first quarter of 2017.12

The machine first approach is based on using automation, AI, and other advanced technologies to enhance the work of people, not to displace them. Indeed, much work based on manual labor will be automated. But that labor needs to be assigned to the new jobs and tasks of a company, which we believe will be bountiful.

Unlike in previous eras, where the brute force of steam engines and gasoline motors replaced human muscle power, the Business 4.0 era will be defined by the combination of human ingenuity with machine speed and power. Humans will be the guiding force helping the machines make good judgments and respond to novel situations.

Humans and machines are already delivering impressive results in the field of robotic surgery. The machine’s precision reduces complications and enables faster recovery, but it is human intelligence that still guides the robotic arm.13

More broadly, human-machine collaboration allows machines to take on dangerous, unpleasant, or monotonous tasks, while freeing humans to concentrate on the creative, intellectual work that produces higher-value products and services. While humans performing monotonous tasks may make mistakes due to boredom and a lack of concentration, regularly-maintained machines can continue to perform the same functions at the same high quality day after day on an indefinite basis.

Even the media industry has seized opportunities to leverage the power of AI. The Associated Press news wire agency has used AI to automate the production of simple, repetitive news stories on topics such as corporate earnings reports and the results of ball games, freeing reporters to focus on more complex and interesting articles.

We see many opportunities for companies to improve and replace laborious and often dangerous manual work, for example:

In manufacturing, machines can take on physically difficult or dangerous tasks while humans assume monitoring and control responsibilities. This division of labor can dramatically reduce health and safety risks for the human workforce while improving productivity. As machines get more advanced and adaptable, they can be reconfigured on the fly to produce new or redesigned products. This human-machine collaboration should dramatically accelerate the speed with which companies can bring products to market, delivering major ROI benefits to companies that can seize first-mover advantages. 

Human-machine collaboration should dramatically accelerate the speed with which companies bring products to market.  

In accounts receivable, machine-human collaboration can help solve the problem of rampant late payments. In 2017, nearly 49% of B2B invoices in the U.S.,Mexico, Canada and Brazil were overdue.15 Machine learning can automate what is currently a highly labor-intensive process of matching invoices with purchase orders. Machines can learn to handle receipts and invoices according to each company’s collection policies. Algorithms can be written that classify receivables and prioritize accounts for collection based on variables such as customer history and invoice data. By extrapolating credit-risk rules from large data sets, machine learning can accurately spot warning signs of credit/payment problems and prioritize those accounts for action.

In warehousing and distribution, robots and other smart machines (including drones) can make warehouses run more efficiently and streamline delivery processes.


In IT departments, automated software distribution, deployment, and patch updates can liberate developers and engineers to focus on developing new products and services.



Barriers and Benefits

Every major technological change is met with resistance. The machine first delivery model will be no different.

Companies will need to invest in reskilling and retraining employees. Retraining efforts should focus on teaching people the skills they will need to perform their new jobs in collaboration with intelligent machines. The IT function must transition from a reactive to a proactive role. Instead of fulfilling orders and meeting demand, the IT function must be prepared to provide best-practice advice and guidance to colleagues in each functional area.

At the same time, companies should look for opportunities to have humans leverage their domain knowledge to help machines get better at their tasks and make automated systems more useful.

The machine first delivery model will be disruptive, but it will not—if correctly implemented— disrupt the large employee base in most industries. Instead, it will make their own work lives more human, not less.


1Cummins press release, March 7, 2018. Accessed August 16, 2018.

2Cummins press release, March 6, 2015. Accessed August 16, 2018.

3The Wall Street Journal, Aug. 2, 2018. Accessed August 16, 2018. 1533255443?mod=searchresults&page=1&pos=1

4Geekwire, Feb. 8, 2018. Accessed Aug. 16, 2018.

5 TechCrunch, Aug. 6, 2018. Accessed Aug. 16, 2018.

6 Zillow webpage, accessed Aug. 16, 2018.

7 Washington Post, Feb. 23, 2017. Accessed August 16, 2018. 2017/02/22/9c6bf5fc-d1f5-11e6-a783-cd3fa950f2fd_story.html?utm_term=.9929ea7ddfd3

8McKinsey & Co., "Advanced analytics: Nine insights from the C-Suite," accessed August 15, 2018, at advanced-analytics-nine-insights-from-the-c-suite

9Sloan Management Review, "Using Analytics to Improve Customer Engagement," accessed August 15, 2018, at engagement/

10"Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation," McKinsey Global Institute, December 2017, accessed August 12, 2018 at media/McKinsey/Featured%20Insights/Future%20of%20Organizations/What%20the%20future%20of%20work%20will%20mean%20for%20jobs%20skills%20and%20wages/ MGI-Jobs-Lost-Jobs-Gained-Report-December-6-2017.ashx

11"How to Land a Job at Amazon," CNBC, July 26, 2018, accessed August 12, 2018 at

12"Amazon headcount drops by 2,900 employees in Q1 — first decline in nine years," GeekWire, April 16, 2018, accessed August 12, 2018 at amazon-headcount-drops-2900-employees-q1-first-decline-nearly-decade/

1313 Robotic Surgery, Mayo Clinic, accessed August 14, 2018 at

14“The Future of Augmented Journalism—A Guide for Newsrooms in the Age of Smart Machines,” Associated Press, accessed August 14, 2018 at images/ap_insights_the_future_of_augmented_journalism.pdf

15“Payment Practices Barometer Americas 2017,” Atradius, September 12, 2017, accessed August 14, 2018 at 2017.html

About the author(s)

Thank you for downloading