A Guide to Building Change Resilience in the Age of AI

Harvard Business Review

2 days ago

There's near universal consensus that AI will fundamentally change how business is done, yet most organizations have not yet seen a substantive impact from their AI efforts. A BCG Global Survey of 1,000 CXOs in more than 20 sectors reports that just 26% of organizations have achieved value from AI, realizing an average cost savings of 45% and 60% higher revenue growth compared to peer firms.
Why such disappointing results? The survey found that among the many challenges organizations face in implementing AI initiatives, 70% are related to people and processes. While it's true that organizations face additional technical barriers such as poor data quality, integration complexity, or infrastructure costs, our collective experience working across hundreds of companies comports with the study's finding: The primary obstacle is the ability of companies to adapt, reinvent, and scale new ways of working. We call this change resilience.
Why Change Resilience Is Scarce
In the past, organizational transformation was episodic. You modernized your systems, trained your people, and operated in a stable environment until the next wave of disruption hit. But now AI is advancing at a pace that far exceeds most organizations' ability to adapt, and the change is unrelenting.
Business leaders find it harder to anchor AI transformation in a traditional roadmap or use conventional means to drive change-management initiatives. Five-year strategies no longer hold. Annual planning cycles can't keep up. Traditional financial, risk, and legal controls lag further and further behind the onset of new risk types. Static operating models become liabilities. Even the newer ways of working, like agile methodologies, broadly adopted during the rise of the software era, aren't sufficient. To compete in this volatile environment, leaders need to embrace continual change, otherwise they face irrelevance from inaction or burnout from chasing shiny objects.
Change resilience is the capability that will equip organizations to seize the opportunities and preempt the threats presented by fast-evolving technology. It's an enterprise‑wide reflex that converts continual disruption into repeatable learning loops that create value. It uses three muscles:
Sensing, or the ability to pick up weak technological, competitive, or societal signals early
Rewiring, or the capacity to redeploy talent, data, capital, and decision rights in days or weeks, not fiscal quarters
Lock‑in, or the discipline to codify what a team learns (in process, code, or policy) so the next initiative starts from a higher baseline instead of reinventing the wheel
Together, these muscles can keep an organization's metabolism in step with AI's rapid advances.
Shopify offers a compelling example of change resilience in action. Rather than layering AI on top of existing operations, the company continuously rewires itself to stay ahead of what's next. In 2023, Shopify made the bold decision to spin off its entire logistics arm, one it had spent years building, to refocus on product innovation. This reset enabled Shopify to rapidly launch AI-native features like Sidekick, an embedded assistant for entrepreneurs that helps with everything from marketing copy to sales insights. By shedding complexity and codifying learnings from past pivots, Shopify unlocked speed and focus, allowing it to serve more than a million businesses with tools that reflect the evolving expectations of digital commerce. Its ability to sense, rewire, and lock in new ways of working positions it not just as an adopter of AI, but as a company continuously reshaping itself to thrive in the AI era.
To understand how change resilient your organization is, ask yourself:
Can employees be redeployed to fast-moving, high-priority initiatives to respond to changes in technological capabilities without the need to overhaul budgets or org charts?
If a team member had an idea today, do they have the motivation, access, tools, and support to start experimenting?
When an experiment shows potential is there a clear path for scaling and embedding it across the business?
Is failure treated as a learning opportunity and openly shared to improve the next attempt?
If you can't answer 'yes' to most of these questions, then your organization does not yet have the change resilience required to turn your AI strategy into durable performance gains.
A Five-Step Playbook for Strengthening Change Resilience
Here are the steps you need to take as a leader and organization to improve your change resilience:
1. Learn: Understand the toolsets, mindsets, and skill sets.
To surface weak spots in toolsets, mindsets, and skill sets, it's important to get your employees to engage with AI and launch experiments. Use these experiments to develop an intuition for how your company might reimagine its processes while also identifying and eliminating cultural barriers that penalizes failed pilots or technical bottlenecks that makes data access a month-long ticket.
Accenture began by encouraging every function (from sales to HR) to build micro-apps that solved a single pain point. Within 10 months this 'sandbox' approach produced 300 generative-AI apps, most of them lightweight utilities such as a proposal draft buddy or meeting-note summarizer. Because each app is owned by the team that built it, employees see immediately how AI reshapes their day-to-day work, shifting the culture from passive adoption to active experimentation.
To fuel participation, Accenture is training 250,000 employees in gen AI skills and giving every learner a safe data playground. Early analysis shows these micro-bets matter: Gen AI is already saving 12% of working hours and boosting output quality by 8.5%, creating momentum for larger transformations.
2. Do: Launch targeted interventions.
Address each gap in change resilience with the lightest-weight move that can generate momentum in weeks, not quarters. If the culture shies away from risk, introduce 'micro-bets': 10-day experiments with a public celebration ritual for learnings, not outcomes. Where skills lag, run cohort-based sprints that pair domain experts with data scientists to ship a working AI concept by sprint's end. The product becomes both a capability and a proof of possibility. If tool sets are the drag, deploy a self-serve data playground or low-code workflow builder so teams can test ideas. When a tactic moves the needle, codify it into playbooks, reusable code, amended policies, and broadcast the template company-wide.
Singapore-based DBS Bank created a monthly 'north star & feedback' ritual that flags cultural, skill, and tooling frictions and then assigns cross-functional 'mini-squads' to attack the biggest challenge. One early scan revealed manual hand-offs that were slowing loan approvals. Within weeks a new AI credit-assessment workflow was live, which now processes around 380,000 lending applications a year and cuts manual work by 85%. Similar micro-interventions have seeded more than 800 production AI models across 350 use cases, generating an estimated US $563 million in economic value in 2024 alone. Each successful fix is codified into a bank-wide playbook through DBS's digital academy, ensuring that every cycle of experimentation leaves the organization measurably more change-resilient than the last.
3. Imagine: Challenge your team to start fresh.
Don't modernize the old operating model; invent a new one. The functions of the future will not look like the work functions of today. AI-enabled organizations have entirely new roles, workflows, and value propositions.
Moderna embodied this creativity when it merged its technology and human resources departments into a single function, aiming to redefine work responsibilities by distinguishing between tasks best suited to humans and those that can be automated. This strategic move, influenced by Moderna's partnership with OpenAI, led to the creation of more than 3,000 customized AI agents for various business functions, including clinical trials and HR operations, fundamentally modernizing the workplace dynamics and roles of HR and technology.
4. Act: Embrace ongoing cycles of measurement, learning, and re-investing
Don't get stuck in the trough of disillusionment. Every wave of technology comes with both inflated expectations and real strategic potential, so it's important to move early, learn fast, and keep going.
P&G has approached change resilience as an ongoing capability, building momentum across toolsets, skillsets, and mindsets with measurable outcomes. Its custom Generative AI platform, ChatPG, now has more than 30,000 employees onboard and supports more than 35 production use cases. In marketing, it has cut concept-testing cycles from months to days, dramatically reducing costs. In supply-chain operations, pilots combining AI with plant-floor sensors have already enabled fully autonomous shifts at Gillette facilities, part of a broader plan that the CFO estimates could unlock $2 billion in productivity gains. On the skills front, the company rolled out AI reskilling, where an initial cohort of 200 employees earned more than 4,400 badges and nearly 90 certifications, applying their learning to dozens of digital initiatives. These metrics allow P&G to tie upskilling directly to business impact and optimize its learning investments accordingly.
Culturally, P&G reinforces a growth mindset through its ' School of P&G,' which blends formal training (10%), mentorship (20%), and on-the-job experience (70%). It further personalizes learning using AI that recommends content and pathways based on individual goals and behavior, an approach that has lifted engagement scores in internal surveys. By integrating bold bets, fast learning loops, and targeted reinvestment, P&G is converting AI experimentation into enterprise-wide performance gains.
5. Care: Put human wellbeing at the center of change
Rapid change can exhaust even the most capable workforce. Without deliberate attention to wellbeing, enthusiasm for new technology quickly turns into fatigue and resistance. The care muscle therefore focuses on creating psychological safety, monitoring sentiment in real time, and giving people the resources (time, coaching, and flexibility) they need to stay healthy and engaged while they learn new ways of working. When leaders treat wellbeing data with the same rigor as financial metrics, they not only protect their people but also accelerate adoption of the very innovations that drive competitive advantage.
Cisco shows how meeting employees' physical, mental, and social needs can accelerate rather than distract from digital reinvention. In 2024, 84% of its workforce logged a combined 2.3 million team check-ins, giving leaders a real-time read on sentiment and workload. At the same time, Cisco embedded an AI-powered 'WellNest' bot to serve up personalized resources for physical, mental, financial, and social wellbeing. These holistic supports have kept engagement high while the company scales AI pilots across the business, proof that caring for people is a prerequisite to sustained, resilient transformation.
. . .
When leaders embrace change resilience, they unlock a flywheel of continuous adaptation that makes each AI experiment faster, each reinvention easier, and each success more scalable. Organizations like Shopify, DBS, Moderna, and P&G show what's possible when change becomes a muscle, not a reaction. In contrast, those who treat AI as a one-time upgrade risk falling into the trap: using new tools to preserve old models. In today's AI era, the divide between early movers and fast followers is widening into a chasm. The difference is not access to technology. It's the courage and discipline to keep changing.