COMPETING IN THE AGE OF AI: Strategy & Leadership When Algorithms And Networks Run The World (Marco Iansiti and Karim Lakhani)

HBR Press has a unique record of publishing seminal books written by Harvard Business School’s academics.
This review is of one in a line of three books books released roughly 20 years apart that have documented fundamental shifts in the way companies compete in their markets, deliver unprecedented value to their customers, and becomine phenomenally successful global companies. In the 1980s, Michael Porter published the first from his now famous Competition trilogy. “COMPETITIVE ADVANTAGE”, changed the way we view strategy. His later follow-on “Competitive Strategy” changed the way we formulate strategy. In the 2000s, Mike Cusumano’s book “COMPETING ON INTERNET TIME” provided a penetrating analysis of strategy-making and product innovation in the dynamic markets of commercial internet. Microsoft was a central player featured in this book, as a behemoth caught on the wrong foot by an emerging internet giant of the time. In 2020, Marco Iansiti & Karim Lakhani have again documented a fundamental shift in the way companies create value with “COMPETING IN THE AGE OF AI”. Its also of coincidental interest for us to note that Microsoft is again featured here – this time as a completely reinvented company that is writing the rules of market competition in the age of Artificial Intelligence.

What is a AI-driven Company?

Iansiti and Lakhani present a framework for rethinking how companies function – redesigning both their business and operating models. They describe how the core of the digital firm works with decisions powered by software, data, and algorithms, which they call the AI factory, and its implications for management and strategy.

Using a series of interesting case studies about companies that are household names, and a few that aren’t yet in the United States, the authors point to how AI-driven processes are vastly more scalable than traditional processes, allow massive scope increase, enabling companies to straddle industry boundaries, and create powerful opportunities for learning–to drive ever more accurate, complex, and sophisticated predictions.

How are these companies different from traditional global corporations we have seen operating so far?

For one, these companies ensure there are no workers in its “critical path” of value delivery, leveraging an approach of AI-driven firm’s operating activities. At the core of this new kind of firm is a decision factory, the “AI factory.” In all cases digital decision factories handle some of the most critical processes and operating decisions. Software makes up the core of the firm, while humans are moved to the edge.

How does this approach create new competitive advantage?

The more customers you have, the more data you can gather, and that data, when analyzed with machine-learning tools, allows you to offer a better product that attracts more customers. You can then collect even more data and eventually marginalize your competitors in the same way that businesses with sizable network effects do. Gathering customer information and using it to make better products and services is an age-old strategy, but the process used to be slow, limited in scope, and difficult to scale up. That changed dramatically with the advent of the cloud and new technologies that allow firms to quickly process and make sense of vast amounts of data. Internet-connected products and services can now directly collect information on customers, including their personal details, search behavior, choices of content, communications, social media posts, GPS location, and usage patterns. After machine-learning algorithms analyze this “digital exhaust,” a company’s offerings can be automatically adjusted to reflect the findings and even tailored to individuals.

Scale, scope, and learning have come to be considered the essential drivers of a firm’s operating performance. In traditional operating models, scale inevitably reaches a point at which it delivers diminishing returns. Rather than relying on traditional business processes operated by workers, managers, process engineers, supervisors, or customer service representatives, the value we get is served up by algorithms.

Organizations driven by AI at their core excel at connecting businesses, aggregating the data that flows among them, and extracting its value through analytics and AI will have the upper hand. Traditional network effects and AI-driven learning curves will reinforce each other, multiplying each other’s impact.

Iansiti & Karimi pont to a Chinese company, Ant Financial, as an interesting case study to illustrate the power of learning-driven scalability. Spun out of Alibaba, Ant Financial uses artificial intelligence and data from Alipay—its core mobile-payments platform—to run an extraordinary variety of businesses, including consumer lending, money market funds, wealth management, health insurance, credit-rating services, and even an online game that encourages people to reduce their carbon footprint. The company serves more than 10 times as many customers as the largest U.S. banks—with less than one-tenth the number of employees. At its last round of funding, in 2018, it had a valuation of $150 billion—almost half that of JPMorgan Chase, the world’s most valuable financial-services company.

The authors also claim many other firms have begun the journey towards transitioning to AI-driven operating model. Many—including Nordstrom, Vodafone, Comcast, and Visa—had already made important inroads, digitizing and redesigning key components of their operating models and developing sophisticated data platforms and AI capabilities. According to them, you don’t have to be a software start-up to digitize critical elements of your business—but you do have to confront silos and fragmented legacy systems, add capabilities, and retool your culture.

Fidelity Investments is another company in using AI in delivering financial services by enabling processes in important areas, such as customer service, customer insights, and investment recommendations. Its AI initiatives build on a multiyear effort to integrate data assets into one digital core and redesign the organization around it.

What happens when AI-driven companies confront traditional companies in the marketplace?

Iansiti and Lakhani explain how “collisions” between AI-driven/digital and traditional/analog firms are reshaping industries, altering the course of our economy, and forcing traditional companies to rearchitect their operating models. They also point out one doesn’t have to be a software start-up to digitize critical elements of your business—but you do have to confront silos and fragmented legacy systems, add capabilities, and retool your culture.

It is a capability that larger mature companies could bring in as well by re-architecting the firm’s organization and operating model. For a long time, the involvement of IT as an enterprise enabler led to the creation of enterprise systems which often reinforced organizational silos and the divisions across functions and products, causing a fragmentation of data meant to create logical linkages piecing together an entire 360 customer view. The new firm aggregates the firms data assets by connecting federated data stores to enabling wider sharing through the extensive use of application programming interfaces and data catalogs to help create predictive models. These models in turn are prepared for deployment via hypothesis driven experiments rolled out across select samples of customers at scale and using learnings from these experiments to create rapid product/service prototypes that incorporate the insights gained from hypothesis testing. The outcome from this deployment approach has resulted in products that have received mass adoption on a global scale literally overnight, that in turn opened a gold-mine of new usage data invaluable to fine-tuning this new product while in flight.

However, in order to achieve this, companies must change their approach to formulating strategy as well.

Instead of focusing on industry analysis and on the management of companies’ internal resources, strategy needs to focus on the connections they could create across industries and the flow of data through the networks the firms use. Network analysis is an essential element of strategy development for competing in a AI world. However to be clear, network effects produce little value before they reach critical mass, and most newly applied algorithms suffer from a “cold start” before acquiring adequate data.

Finally in conclusion, the authors outline a few considerations business leaders and regulators must confront in this new world. Algorithmic bias, data security and privacy – as well as platform control and equity – are new challenges and responsibilities leaders of both digital and traditional firms need to address. “Our new meta is generating enormous opportunity, as evidenced by economic growth. But its also leaving us struggling to understand the full implications of the new rules, dealing with a range of new problems, and coping with an increasingly complex consequences”.


An organization’s future success depends on their decision makers’ ability to anticipate changes and disruptions in the marketplace. But how do you get information about tomorrow today? How can your decisions today account for
tomorrow’s uncertainty. Sensing the faint signals that provide a glimpse of our evolving future is never an easy task. Nor is it foolproof. An even trickier job is sorting out false background radiations masquerading as information,
from the true signals.

There has been a mushrooming of offerings in the past decades built around on divining the future. Most have evolved as offshoots of academic research groups. Stanford Research Institute (later SRI International) is a pioneer in seeing into the future and one reputed for developing some of core futuristic technologies that have today become common language and embedded into our daily lives.

Martin Schwirn, author of a new book “Small Data, Big Disruptions” works for Strategic Business Insights. SBI is the   former Business Intelligence division of SRI International that has worked with clients on opportunities and change  since 1958 to help  identify and map new opportunities based on emerging technology and market insights. SBI focuses on seizing information opportunities that exist between the moment you sense early signs of emerging disruptions and when it becomes public knowledge to capitalize on.

He introduces a brand of the futuring methodology, called Scanning.

Scanning offers a four-step process for capturing and analyzing information from a company’s external environment, helping decision makers foresee coming changes in the marketplace:

Filter vital information from an avalanche of data.
Identify what matters most to you and your organization.
Prioritize the crucial changes that will shape tomorrow’s marketplace.
Initiate strategies that move you from vulnerability to preparedness.

Scanning uses brushstrokes, not to create a perfect picture of tomorrow, but to develop insights into the way
circumstances, technologies, consumers and businesses might align to create markets. It makes the initial sketch, which evolves into a painting with emerging data on certainty.

Becoming aware of the world is the first step.

Topics and events that haven’t entered our active conscious radar, and yet have been picked up on our keen peripheral vision, have the potential to manifest themselves as factors of future uncertainty within the present business planning context of organizations. A Scanning type method often helps identify them.

The next challenge is assigning meaning to changes and disruptions.

This is quite another thing.  Identifying a laundry list of potential future trends may create quite a pile of signals. That makes the step vulnerable to the same challenges of poor signal-to-noise ratio as we have experienced in tuning AM/FM broadcasts on our radios in the past. More recently, this is the core challenge data scientists aim to resolve in applications related to deployment of artificial intelligence technologies. The task of finding those signals that matter to our organizational vision and future success must be taken as seriously as doing the baseline environmental analysis.

Equally important are the choices made to filter out background noise.

Making good choices to discard, clarify or sharpen signals efficiently and lead us to a fruitful visioning session within a period of deliberation. On the other hand, lingering too long on ambiguous signals (or being lost looking for that needle within a large haystack) risks diluting the goal of convergence, and prolonging the period of confusion among senior management, seriously jeopardizes or even leads to postponement of critical strategic investment decisions.

It is this efficiency of separating the true signal from background noise that proves the difference between an effective futuring process and a less successful one.

Avoiding the trap of false positive signals
A specific early signal frrom the past provides interesting humourous anecdotal evidence of information to watch for as one sifts through positively evolving signals.
Criminals are often early adopters of new technology, since they have a strong incentive to be innovative in that way. But that doesn’t mean the technology itself is the problem; new discoveries and their applications are inevitable. In the late ’80s, criminals were early adopters of beepers in the 1980s. Ths snippet from a July 1988 Washington Post report describes it –
” When a drug dealer is in trouble, he sometimes dials 911. But he isn’t trying to reach the police. Instead, this message is sent to a drug courier wearing a beeper that displays messages dialed from a phone: 911 means the police are closing in.
About 6.5 million beepers are in use in the country, according to officials, although it is difficult to estimate what percentage is used for drug trafficking. Federal narcotics agents estimate that at least 90 percent of drug dealers use them. U.S. Drug Enforcement Administration officials said that beepers, which have been used by bookies and cigarette smugglers, were introduced in the drug market about five years ago by Colombian cocaine organizations.
Although paging devices, or beepers, have grown in popularity throughout the labor force – doctors, delivery people and journalists often use them, they also have become a staple in the drug business, posing fresh problems for law enforcement and threatening to tarnish the image of a booming high-tech industry.”

Doing environmental scanning the right way identifies key mosaic pieces that point to evolutionary dynamics of a certain future.

The author appears challenged to provide a clear path forward here. This book is replete with examples repurposed from well-known past business school cases and those used in other contexts. As with most “disruption” examples in the book (visible post facto), using them to make the case for a method to evaluate futures doesn’t quite convince the reader of its value, that future outcome being already well known widely.

On the other hand the author’s situation is quite understandable. It is certainly a challenge to demonstrate the power of this futuring process by identifying key uncertain trend signals of the future, and then observing them play out as a certain future, all within the same time-frame of the present. Therein lies the absurdity. How does one navigate this?

Identifying cause-effect relationships between narratives

One option to do this may be by identifying and fleshing out cause-effect or correlational relationships between
identified narratives. This brings us to the point of using broader narratives to describe each signal rather than short specific statement of facts. To acknowledge that these signals to the future are still ambiguous within
our present time-frame. Also to acknowledge that some or many of these signals may well not play out as we expected, or even turn out to be downright false.

Identifying these cause-effect relationships require that narratives (while ambiguous representations of the future) must be worded precisely enough so as to separate them from others in the box. As things progress, futuring workshops often see groupings of narratives over time, as new information presents itself to emphasize such relationships. Another benefit of mapping such cause-effect relationships is that it forces futuring teams to deepen their quality of research and insights into each signal/narrative identified, and leverage the knowledge & experience of subject matter experts.

Martin Schwirn shares other important insights in his book.

He emphasizes the importance of casting a wide net while gathering raw materials before filtering them and to discern the faint signals to the future. Having good peripheral research vision is a very strong skill-set here. There is place for subject specialists here, but there is an equally important leading role that generalists must play in helping connect information-dots and developing the most effective narratives, during the process.

The difference between complex futures and contrasting futures.

In helping us make the case for this futuring methodology, he points to some important differences between complex futures and contrasting futures. I found this quite interesting.

While one potentially lends itself to big data approaches for resolution, the other necessarily relies on small data. While one is about perfecting the outcome as it existed in the past – repeatable, predictable, actionable in the short term, where looking at huge historical data sets certainly help clarify signals from the noise very efficiently and cost effectively, and with an accuracy and breadth the human brain could not handle. The other is about peering into a foggy road ahead, armed with robust conceptual aids to sharpen focus, and to put just enough boundaries around the scope of vision to avoid creating scenarios that distract from the planned outcome of futuring sessions.

He points out how big data is not a suitable approach to evaluate potential futures, when dealing with forward looking uncertainty. Why use of big data has the missed the signal to many profound environmental changes that affect organizations.

The author also rightly emphasizes the criticality of involving multidisciplinary teams. Scanning requires organizations to lend their best informed experts to participate in the generation and evaluation of futures narratives. There is place for subject specialists here, but there is an equally important leading role that generalists must play in helping connect information-dots and developing the most effective narratives, during the process.

Evaluate Trend Signals on Impact and Emergence Dimensions

This is a significant insight. Generating many early signals and narratives naturally leads us to the need to manage them in a hierarchy in order to make sense and prioritize them for requirement of the organization. He proposes an objective way to do so using clear metrics and criteria. While Impact may be easier to assess in somewhat quantitative terms, using known financial or operational measures. Qualifying the Emergence requires an assessment of its relative uncertainty in playing out as well as the period over which this may happen, a qualitative skill.  

We also observe the book left a few areas wanting.

It makes passing references to some evolving significant technology trends that are in midst of creating the
most profound impacts on the way we live, work, manage our business and (in some cases) even how we think. The potential of technologies such as the internet-of-things, artificial intelligence, robotics, intelligent cities, cryptocurrency, virtual reality, gaming, evolution of metaverses, and many others that are still playing out their disruptive potential.

However the reader does not quite come away with any significant understanding or insight into how and why the potential of these technologies are important narratives in our presently evolving futures. Barring passing mentions of the terms, the book does not appear to have leveraged any deep research or subject matter expertise for interpreting those trends.

[ Author’s note: I may carry a personal bias here. My firm ArcInsight Partners specializes in studying technology trends such as industrial internet of things, evolution of intelligent cities, artificial intelligence applications and autonomous manufacturing. It carries out deep industry research and supports clients in making critical high value high impact business decisions that concern its mediun to long term future – assigning relevant market trends and weighting them for business units, strategies for taking new products/services to market, identifying acquisition targets, strategies for monetizing new business models, even providing interim strategic leadership roles for their internal planning teams. ]

The book stops shy of examining the profound impact on our society and our environment brought on by fast
evolving technology. What are the imperatives that global shifts such as the global socio-political protest movements and rapid climate change events bring into the futuring process.

It misses the opportunity to walk the reader through the experience of a futuring session, how some of today’s strong contemporary trends that are sure to feed into signals for our still evolving future. , how an assessment of narratives are organized to paint a future. Perhaps even make few bold predictions about the future, standing in the present. Knowing well that many do not play out as expected.

It makes a weak attempt to acknowledge the possibility that companies act to shape futures (as it emerges), as much they adapt to them. Large successful transnational corporations and mega multilateral organizations cannot afford to rely on an exercise of future-generation tools as their only risk management tool. They are the most vulnerable to generational shifts, and consequently deploy a plethora of tools and risk-mitigation techniques to manage the future. Successful companies are increasingly grabbing a bigger share of the future signals by investing in corporate venture capital funds, in order to turn those signals into a portfolio of experimental businesses and venture investments some of whom it expects to fail. The failures would be their greatest success in generating verifiable data to identify and filter signals.

Where should futuring teams reside within a complex enterprise?

How does one know a Scanning session has been a success?
How does a business enterprise make the case to have Futuring as a separate organizational entity?
What milestones indicate progress towards understanding the future?

These remains unaddressed question. Without some clarity on these questions, the task of futuring inevitably returns to its small place within the corporate strategy fold – and as strategists go, they being sticklers for evidence and spreadsheet quantification. The very things that may be stumbling blocks for Scanning as a stand alone methodology.

We look forward to Martin Shwirn’s follow-on discussion that goes deeper than cursory explorations of a few future narratives identified in the book, especially some those that are yet to play out entirely. Perhaps it would delve into specific anecdotes and cases the author experienced directly as practitioner. These we believe could certainly make an exciting read journey of exploring hidden signals to the future all around us, and make a few bold predictions of how they might unfold within our lifetimes.


Amazon case study – from book retailer to AWS cloud infrastructure provider to space-travel services. Was it Futuring, or a series of calculated risk-