Key Takeaways
Transform people’s behavior, not merely the technology they touch
Technology needs new habits. Digital tools create little value when people use them to repeat old routines. At Singapore’s DBS Bank, the 2009 transformation began with the recognition that technology had to support a different kind of bank: more innovative, more responsive, and more like a technology company. DBS brought operations in-house, moved many applications to the cloud, and reorganized technology and business leaders around shared platforms and goals.
Culture is the delivery system. By 2016, Euromoney named DBS the world’s best digital bank; in 2019, it received three major global bank-of-the-year awards. The story’s lesson is not that technology was irrelevant, but that structures, incentives, and daily behavior had to change for technology to matter.
The DBS example makes digital transformation a problem of organizational behavior, not procurement. That aligns with sociotechnical thinking: tools and social arrangements shape one another, so changing one side alone often preserves the old system in a shinier form. The bank’s shared platform accountability also addresses a familiar coordination problem: when business and IT own separate goals, each can blame the other for outages or delay. A caution is that successful case studies can make change appear more orderly than it feels inside an organization. The practical test is whether decision rights, routines, and measures actually shift, not whether leaders announce a new culture.
Earn the right to reinvent by learning through smaller transformation stages
Maturity accumulates in stages. The book describes a learning curve that moves from modernization, to enterprise-wide transformation, to new business creation. Modernization digitizes existing work, such as customer self-service or automated employee processes. It rarely changes the business model, but it can produce early returns and teach the organization how to connect systems, improve processes, and gain user adoption.
Complexity rises with ambition. Enterprise-wide transformation changes how functions work together, such as integrating customer channels or using connected equipment for predictive maintenance. New business creation goes further, changing how the company earns value, perhaps through subscriptions or ecosystem partnerships. Firms can run initiatives across all three stages at once, but they should not assume that a pilot or a new business model can substitute for the capabilities built through earlier work.
This staged view resembles capability development in fields where expertise depends on repeated practice, not simply a bold target. Modernization can look modest beside a promised new business model, yet it exposes the hidden work of data quality, integration, and adoption. The stages should not become a rigid ladder, however. A company facing an immediate threat may need a protected venture or rapid experiment while foundational work continues. The useful distinction is between sequencing learning and delaying ambition: leaders can pursue future-facing bets, provided they understand which organizational abilities those bets assume and fund the gaps.
Organize around products so customer learning can compound
Products connect teams to value. The flywheel operating model puts multidisciplinary product teams at the center of work. A product can be a physical good, a service, a digital experience, or an internal tool, as long as it solves a user problem. Teams own its ongoing development rather than completing a one-time project and handing it off.
Learning becomes a repeatable loop. Teams gather user feedback and operational data, identify opportunities, run experiments, then build and deploy useful changes. Shared technology platforms provide reusable data, services, and infrastructure, while leaders give teams room to make decisions. A BCG study cited in the book found that companies scaling this digital-first model produced 2.5 times higher total shareholder value from 2018 to 2021 than the comparison group. That result supports the model, though it does not prove the model alone caused the difference.
The flywheel’s strength is that it links customer contact to organizational learning. Product teams can correct a small defect before it hardens into an expensive operating assumption, while platform teams reduce duplicated technical work. This also explains why adopting agile project rituals alone can disappoint: a team may build faster without having authority or evidence to choose what is worth building. The model depends on genuine autonomy and reliable shared infrastructure. Without those, product teams become another label layered over functional silos, and the promised cycle of feedback and improvement stalls.
Make culture concrete by designing the behaviors people repeat
Small mechanisms can shift habits. DBS developed BEANs, meaning behavior enablers, artifacts, and nudges: practical structures that make a desired behavior easier to perform and remember. For meetings, its MOJO routine assigned an owner to set an agenda and a joyful observer to encourage participation and give feedback. The book reports that the approach saved nearly 150,000 employee hours annually, doubled perceived meeting effectiveness, and raised the share of employees who felt they had an equal voice from 40% to 90%.
Find the real blocker. At law firm King & Wood Mallesons, lawyers hesitated to use productivity software partly because efficiency could reduce billable hours, a measure tied to perceived performance. The firm allowed training time to count toward targets and gave additional internal credit for using productivity tools. Changing the incentive addressed a deeper barrier than simply offering software training.
BEANs translate a broad aspiration, such as collaboration, into observable situations and cues. This is close to behavioral design: routines, social expectations, and incentives influence conduct more reliably than slogans alone. The law-firm example is especially revealing because resistance can be rational within the system people inhabit. A tool that improves client value may still threaten an employee’s status or income. These interventions need evaluation, since a visible ritual can become performative or intrusive. The strongest designs make the desired behavior useful, safe, and compatible with the outcomes people are already rewarded for.
Treat AI readiness as a data and work redesign problem
Data preparation is the heavy lift. The authors estimate that roughly 80% of the work of preparing a machine-learning algorithm goes to gathering and cleaning data. A company’s claim that it has years of records does not mean those records are usable: they may be scattered, inconsistent, or missing the details a specific decision requires. Leaders should start with a concrete business problem, then establish what data is available, what must be collected, and who will maintain it.
AI changes routines, too. Readiness also depends on staff training, executive understanding, tolerance for mistakes, and willingness to replace outdated practices. AI can help select projects, improve planning, or surface risks, but it should complement human judgment. Its recommendations need to be understandable and usable within real decisions, rather than added as a technical layer that leaves existing workflows untouched.
The data-first argument counters the tendency to treat AI as a shortcut around operational disorder. In practice, data work often forces organizations to clarify definitions, ownership, and process boundaries, which can be valuable even before sophisticated models arrive. Human-machine collaboration also requires more than trust in an algorithm: users need ways to inspect assumptions, question outputs, and provide corrections. The book’s readiness framing is useful, though a single score cannot settle whether a particular use is safe or worthwhile. Data quality, decision stakes, and the consequences of error vary sharply by application.
Use digital capability to adapt, not just automate the present
Compete through adaptation. One chapter reframes the company as a living system rather than a machine built only for efficient repetition. Digital tools can extend what an organization notices, connect data to rapid decisions, and help people and algorithms contribute different strengths. Netflix’s recommendation system, for example, uses viewing behavior to refine personalized suggestions. The broader aim is to improve how a company senses change, interprets it, imagines responses, and acts with others.
Define the purpose first. The book’s “conscious enterprise” idea similarly points toward organizations that use data and technology to respond to customers and changing conditions, through capabilities such as automated workflows, configurable products, and more responsive supply chains. Yet technology cannot decide what a company ought to value. Purpose, ethics, and the acceptable use of customer data remain human responsibilities.
Thinking of a firm as an adaptive organism shifts the question from “Which process can be automated?” to “What can the organization perceive and learn that it could not before?” That is a richer strategic test, especially when customer preferences or industry boundaries move quickly. It also brings an ethical constraint into the center: greater sensing power can improve service, but it can also enable surveillance or manipulation. The metaphor should not obscure the company’s institutional responsibilities. Unlike biological adaptation, business choices affect workers, customers, and communities who can contest the goals and costs of change.
Invite outsiders to attack your business model before reality does
Disruption hides at the edges. Established firms often overlook threats because they target customers the incumbent considers unimportant or come from outside its industry. A Nightmare Competitor Contest uses a war-game format to challenge those assumptions. Executives work with outsiders, such as customers, start-ups, or experts from other sectors, to invent a plausible digital competitor and examine how it could take value away.
Turn the scenario into a response. The workshop starts by agreeing on the current business model and relevant trends. Participants then build competing disruptor models, select the most threatening, and assess the risks and opportunities. The exercise is designed to reveal weak signals and expand the range of strategies leaders consider. The book also argues that partnerships with start-ups can help established firms explore new technologies and respond to shocks, if both parties share value and collaborate effectively.
The contest is valuable because it makes strategic disagreement productive. Instead of asking whether a threat is likely, participants temporarily inhabit the logic of a competitor and ask what could make the threat work. This resembles scenario planning, but its competitive format can loosen attachment to familiar assumptions. The exercise is only as good as the outsiders’ credibility and the executives’ willingness to hear unwelcome conclusions. A vivid scenario can still be mistaken for a forecast. Its purpose is not prediction, but preparedness: expanding the set of plausible futures and creating options before a crisis removes the time to think.
Expect value creation to shift from pipelines to platforms to protocols
Business structures follow technology. A pipeline creates value through a linear path from producer to customer, as in mass manufacturing and retail. A platform connects many producers and consumers, helping them find one another and transact; more participants can make the service more useful. The book argues that digital connectivity, cloud computing, and data helped platforms scale, while also concentrating market power in a few intermediaries.
Protocols redistribute roles. Blockchain-based protocols can coordinate transactions without one central platform controlling every interaction. Tokens may reward contributors, and digital records can support ownership and verification. The author argues that this could give producers more control over their assets and consumers more control over identity and data. Protocols are not guaranteed to replace platforms: they need infrastructure and workable incentives, and the book expects all three models to coexist where each fits best.
This framework is useful because it treats technology as a change in market organization, not merely a new product category. Each structure solves coordination problems differently: pipelines centralize production, platforms reduce the friction of matching, and protocols distribute some verification and governance. The protocol case is promising but not settled. Token rewards can create speculation, governance can be captured by influential participants, and decentralized systems still depend on infrastructure and rules. The practical question is therefore not whether decentralization is virtuous, but which activities benefit from shared control and which require accountable intermediaries.
Distribute innovation with guardrails and measure outcomes, not activity
Business users can build safely. Citizen development lets employees create applications with low-code or no-code tools, translating their knowledge of work directly into prototypes and solutions. Without IT-approved platforms, training, and oversight, however, these tools can become “shadow IT,” creating security, compliance, and maintenance risks. PMI’s Citizen Development Canvas offers a structure for delivery, capability building, governance, and organizational adoption. Amtrak’s program grew from six participants and 16 applications in 2020 to 478 developers and 4,985 applications by 2023; the company reported savings of $76 million and more than 1.2 million staff hours returned.
Transformation offices need strategic standing. The book’s research finds that mature offices connect projects to executive priorities, build talent, and track outcomes such as customer satisfaction or efficiency, rather than relying only on schedule and budget. A project is not successful merely because it was delivered on time if it failed to create the intended value.
Citizen development and transformation offices address opposite sides of the same organizational challenge: how to expand initiative without losing coherence. Local builders can shorten the distance between a problem and a solution, while governance can protect shared systems and data. Outcome measures help keep both accountable to value rather than activity. The Amtrak figures are striking but should be read as a reported case, not a guaranteed return for every firm. Scaling applications also raises questions about upkeep, duplication, and who owns a tool after its creator changes roles.
Use AI to make sustainability evidence visible, not merely persuasive
ESG information is fragmented. Investors often rely on ratings that disagree because agencies use different definitions, measurements, and company-provided data. The book cites Boohoo, which received strong ratings despite serious allegations about labor conditions in its supply chain. It also notes that financed emissions, produced through lending and investment, can be far larger than a financial institution’s direct emissions and are difficult to track.
AI can widen the evidence base. The proposed process is to harvest information from reports, news, and other sources, organize it into structured data, then analyze its meaning and context. This can help investors monitor developments and compare companies more frequently. But AI cannot by itself settle which standards matter or whether a source is credible. Human judgment and transparent methods remain necessary if automated ESG analysis is to support accountability rather than create another opaque score.
This application shows how AI can improve the information environment around a public-interest problem. Processing unstructured reporting at scale may expose inconsistencies that conventional ratings miss, giving investors and companies earlier signals. Yet automation can reproduce bias in its sources, confuse allegations with verified findings, or make uncertain evidence look precise. The deeper issue is governance: who defines a meaningful sustainability measure, who can challenge it, and how errors are corrected. AI is most useful here as an aid to scrutiny, not as an authority that turns contested judgments into an apparently objective number.
Analysis
This collection is best read as a field guide assembled from many contributors, not as one tightly unified theory. Its recurring claim is that digital transformation is a change in how an organization creates value and learns, with technology as an enabler rather than the transformation itself. The most useful synthesis is a chain: identify a customer or operational problem, build the data and skills needed to address it, give a capable team authority to experiment, and change incentives and governance so successful learning can spread.
The book’s strongest contribution is its attention to the organizational conditions often hidden behind technology programs. Product teams need modular systems and decision rights. AI depends on data quality and redesigned work. New tools require trust, training, and a reason for employees to adopt them. The examples from DBS, Enel, law firms, and Amtrak make these constraints tangible, while the learning-horizon model discourages leaders from mistaking an ambitious target for organizational readiness.
Its breadth is also a limitation. The chapters use different definitions of transformation and report success and failure estimates from different studies, so their numbers should not be treated as one comparable dataset. Some claims about blockchain protocols, AI, and future operating models are projections rather than settled outcomes. A reader should distinguish tested practices, such as starting with a concrete use case, from hypotheses about where markets are heading.
The most practical test for any initiative is therefore threefold: Does it improve a clearly identified outcome for users or the business? Can the organization support adoption with appropriate skills, data, and authority? Is there a credible way to learn from results and adjust? If those conditions are absent, a technology investment risks becoming modernization theater. If they are present, even a modest pilot can teach the organization how to transform at greater scale.
Glossary
Flywheel operating model
Product-led learning systemAn operating model that organizes people and technology around empowered, multidisciplinary product teams. Teams use customer and operational data to identify opportunities, run experiments, and continuously improve products. Shared technology platforms and governance support their work, while decision-making moves closer to the people responsible for delivering user value.
Digital transformation learning curve
Three stages of capabilityA progression in organizational learning that moves through modernization, enterprise-wide transformation, and new business creation. Modernization improves existing processes and builds foundational capabilities. Enterprise-wide change develops the ability to coordinate across functions. New business creation uses that maturity to develop new markets, revenue sources, or business models.
BEANs
Behavior enablers, artifacts, nudgesA method for changing organizational habits through behavior enablers, artifacts, and nudges. Enablers include routines or rules that make a behavior possible, artifacts provide visible reminders, and nudges prompt people at relevant moments. The method starts by identifying a specific desired behavior and the underlying obstacle preventing it.
EPOV model
Motivation chain for behaviorA model of behavioral motivation linking effort, performance, outcome, and valence, or the value a person places on an outcome. People are more likely to change behavior when they believe effort can produce the required performance, that performance will lead to a meaningful outcome, and that the outcome is worth pursuing.
Citizen Development Canvas
Governed low-code adoption frameworkA framework for introducing, managing, and scaling application development by nonprofessional developers using low-code or no-code tools. It addresses project delivery, capability development, the operating model, organizational alignment, and organizational maturity. Its purpose is to enable business-led innovation while keeping IT involved in platform selection, security, and governance.
Nightmare Competitor Contest
Disruption-focused strategy exerciseA workshop in which company leaders and credible outsiders design a hypothetical digital competitor that could disrupt the current business. Teams develop competing business models, select the most threatening scenario, and examine strategic responses. The exercise aims to reveal threats outside the company’s usual industry assumptions, not to predict one certain future.
Antifragility
Gaining from disruptionAn organizational capacity to turn shocks or major changes in market conditions into opportunities for competitive advantage. Unlike basic resilience, which means surviving a shock, antifragility involves identifying new possibilities and adapting the business model to benefit from them. The book presents collaboration with start-ups and other ecosystem partners as one way larger firms can develop this capacity.
Digital Knowing-Doing Gap
Knowledge not yet practicedThe difference between understanding digital concepts and being able to apply them in work. The cited Digital Maturity Index separates people into reacting, embedding, and strategizing stages. Its reported results show many participants had digital knowledge but had not yet developed the skills or routines to incorporate it consistently into practice.
Conscious enterprise
Purposeful, adaptive digital businessAn aspirational model of a digitally transformed organization with stronger abilities to make sense of changing conditions and act in alignment with its purpose. The chapter associates it with capabilities such as customer sensing, automated workflows, adaptable products, modular technology, responsive supply chains, and broader measures of business impact.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.