Briefing #9: The Opposite of Chaos
Deconstructing how real AI transformation happens.
Note: This briefing was originally published on LinkedIn on September 19, 2025. It has been migrated to our new home on Substack to create a complete archive. Multi-format features like video and audio commentary are available for all new briefings published from April 2026 onwards.
The prevailing narrative of AI-driven transformation is one of radical, chaotic disruption. It’s a story of legacy systems being torn out, roles being upended, and organizations struggling to keep pace with a technology that feels both omnipotent and uncontrollable.
But what if that narrative is wrong? What if the most successful transformations aren’t chaotic at all, but are instead models of discipline, rhythm, and systematic learning?
A powerful case study on Singapore’s DBS Bank, detailed in Harvard Business Review, offers a compelling counter-narrative. The bank undertook a massive AI transformation that resulted in over 800 production models, 350 distinct use cases, and an incredible $563 million in economic value in a single year. Yet, they achieved this without the widespread resistance and project failure that defines most AI initiatives.
Their story is crucial because it provides a working blueprint that directly addresses the core reasons most transformations stall. Research confirms that the primary obstacles aren’t technical. HBR data shows that 70% of AI implementation challenges are related to people and process, and a staggering 75% of companies fail to achieve meaningful value from their AI investments.
DBS succeeded where most fail because they didn’t focus on installing technology. They focused on building an organizational engine for change. By deconstructing their approach, we can uncover three universal principles that are applicable to any leader, regardless of industry.
Principle 1: From “Big Bang” to Micro-Interventions
The common approach to transformation is the “Big Bang” — a massive, multi-year program with a huge budget and an army of consultants. This approach is slow, expensive, and creates enormous organizational resistance.
DBS did the opposite. They embraced what can be called micro-interventions.
Instead of a monolithic transformation office, they formed small, agile “mini-squads.” These cross-functional teams were tasked with identifying a single, specific point of operational friction — like the manual hand-offs slowing down their loan approval process — and deploying a targeted solution, often in a matter of weeks.
This approach flips the traditional model on its head. It minimizes risk, accelerates time-to-value, and builds momentum with a series of small, tangible wins. It proves that transformation doesn’t have to be a single, seismic event; it can be a continuous series of deliberate, high-impact adjustments.
Principle 2: From “Pilots” to “Learning Loops”
The most common failure pattern in corporate innovation is “Pilot Purgatory”—the endless cycle of successful pilots that go nowhere. DBS avoided this trap by replacing isolated experiments with a systematic “change resilience” learning loop, consisting of three phases:
Sensing: The process begins with structured rituals, like monthly feedback sessions, to systematically sense where the most significant friction exists in the business. This isn’t a one-off brainstorming session; it’s a continuous, rhythmic process of discovery.
Rewiring: This is the action phase, where a “mini-squad” is deployed to rewire the broken process with a new solution, like the AI-powered credit assessment workflow that now handles 380,000 applications a year. This is where most companies stop. They prove the technology works and declare the pilot a success.
Lock-in: This is the crucial, and most often ignored, final step. DBS systematically locks in the knowledge from every successful intervention. The solution isn’t just a piece of code; it’s codified into playbooks and training materials for their digital academy.
This “Lock-in” phase is the antidote to Pilot Purgatory. It ensures that every success raises the entire organization’s baseline capability. It transforms a one-off project into a compounding institutional asset.
Principle 3: From “Tech-First” to “People-Centric”
Finally, the DBS model reinforces a truth that is consistently validated by research from institutions like MIT Sloan: successful AI adoption is a managerial and cultural challenge, not a technical one.
Their “mini-squads” were not just teams of data scientists; they were cross-functional units that included domain experts from the business. Their monthly “sensing” rituals created a rapid feedback loop that ensured employee concerns were heard and addressed. This people-centric approach builds the trust and psychological safety required for genuine transformation. It reframes AI as a tool to augment and strengthen the workforce, not replace it — a perspective that is critical for avoiding the internal resistance that sabotages so many initiatives.
The lesson from DBS is profound. The goal of a leader isn’t to oversee a chaotic, unpredictable revolution. It’s to build a calm, predictable, and highly effective engine for change. The $563 million in value isn’t the real story. The story is the repeatable, scalable system that produced it. That is a blueprint any leader can, and should, follow.