By Mark Blackwell, Arkaro
Most organisations are asking the wrong question about AI. They are treating it as a technology to be managed rather than a capability that should serve their business goals. The result is an executive vacuum: the top floor sees the potential enthusiastically, whilst the middle of the organisation is paralysed on the how.
In a conversation with Charlene Li on the Arkaro Insights podcast, we explored why most AI transformations stall — and what leaders can do about it in 90 days. Charlene is one of the world’s foremost experts on disruptive transformation, adviser to 49 of the Fortune 100, and author of six books including her latest, Winning with AI: The 90-Day Blueprint for Success, co-authored with Katia Walsh, Ph.D.
Her central thesis is both simple and unsettling: growth creates disruption, and the leaders who thrive are those who run towards it rather than wait for it to settle.
Executive Summary
Key insights from Charlene Li’s conversation with Mark Blackwell:
You do not need an AI strategy. You need a business strategy that is enabled by AI.
The biggest barrier to AI adoption is not the technology — it is people’s need to protect their identity and sense of purpose.
Data will never be perfectly clean. Start with minimum viable data quality against your highest-priority applications.
Speed is the new competitive moat. Annual planning cycles are a structural obstacle to AI transformation.
The Playground Paradox: clear boundaries do not constrain innovation — they enable it.
HR is the most overlooked seat at the transformation table. If you are changing how people work, you need HR at the centre.
The CEO cannot delegate AI transformation. They must be visible, learning, and experimenting.
90 days is enough time to move from AI curiosity to meaningful implementation — if you focus on your biggest business problem first.
Stop Building an AI Strategy
One of the first and most important distinctions Charlene draws is this: you do not need an AI strategy. What you need is a business strategy that AI enables.
In too many organisations, the CEO thinks about the business over here and AI over there as a separate thing to be dealt with. What is the ROI of AI? How do we manage AI risk? These are the wrong questions. The right question is: how can AI help us do what we are already trying to do — faster, more efficiently, more cheaply, and perhaps in ways we could not have attempted before?
This shift in framing is not semantic. It changes who leads the conversation, how success is measured, and crucially, how much urgency the organisation feels. If AI is a technology initiative, it sits in IT. If AI is a business capability, it sits with the CEO.
Reimagining Business Models: Three Companies Doing It Differently
The most powerful part of Charlene’s work is not the framework — it is the case studies. Three examples from the book illustrate what genuine AI transformation looks like when it is connected to business goals rather than technology novelty.
Konecta: From Cost Pressure to New Services
Konecta is a global business process outsourcer — the kind of call centre company that most people assume will be devastated by AI. Their response was the opposite of what you might expect. Rather than using AI to cut headcount, they used it to raise the quality and capability of their people across the board.
Error rates fell dramatically. Training times decreased. And then the more interesting question emerged: what new services can we now offer that we could never have delivered before? They identified clients — law firms, for instance — who lacked the capacity to deploy AI tooling themselves, and built entirely new service lines around those clients’ unmet needs. A business you would write off became a case study in AI-enabled growth.
Nestlé: Consumer Research at Qualitative Depth and Quantitative Scale
Nestlé used a company called Outset.ai to conduct AI-led consumer interviews. The traditional research trade-off has always been quality versus quantity: rich qualitative insight from a small sample, or statistical reliability from a large one. AI dissolved that trade-off entirely.
Outset.ai conducts spoken interviews at scale. Respondents speak rather than type, in any language, and the AI follows threads, picks up nuances, and probes uncertainty in ways that a scripted survey never could. For the first time, Nestlé could ask genuinely open questions of thousands of people simultaneously and receive insight that was both statistically robust and contextually rich.
Moderna: Putting HR at the Centre of AI Transformation
Perhaps the most structurally radical example in the book. Moderna placed their CHRO, Tracy Franklin, in charge of the entire technology function, including AI. Their entire technology group moved under HR. The logic was straightforward even if the decision was not: transformation is fundamentally about people, not systems. Technology is the enabler. People’s strategy is the core.
The result was a shift in what HR does. Rather than managing headcount and org charts, the function began orchestrating how work flows between humans and AI systems. Performance metrics evolved. Job design changed. Career paths were reconsidered. This is what Charlene means by the superhuman workforce: not AI replacing humans, but humans and AI integrated in ways that allow each to do what they do best.
Who Should Be in Your Transformation Team?
Charlene is direct about the CEO’s role: they cannot delegate AI transformation. This is not about being involved day to day, but about being visible, present, and learning. The CEO is where the buck stops on go or no-go decisions. They are also the most powerful signal sender in the organisation — if they are experimenting with AI openly, including failing and sharing what they are learning, it gives everyone else permission to do the same.
Beyond the CEO, the core transformation team needs five perspectives:
- A technology or AI expert — not necessarily the CIO, who is often focused on administration and risk rather than transformation.
- Someone who deeply understands the customer and how AI will change the experience.
- A business leader who understands revenues, costs, and what success looks like commercially.
- An HR leader who thinks beyond recruitment and risk mitigation to the future of work.
- A communications lead who can create a consistent internal and external narrative.
Eventually, one person needs to own AI adoption — someone for whom it is their entire job to wake up each morning and ask how the organisation is using AI to create value.
The Real Reason People Resist AI
This is where Charlene’s thinking connects directly to the neuroscience of change that Hilary Scarlett has explored so extensively. The biggest resistance to AI is not technical scepticism. It is the threat AI poses to people’s sense of identity and purpose.
A marketing professional who has built their career on crafting beautiful prose is not resisting AI because they misunderstand it. They are resisting because AI appears to obliterate the very thing that defines their professional value. The response to this is not to minimise the threat — it is to reframe what value means. If AI writes the first draft and generates multiple options, the human’s value shifts from blank-page creation to expert judgement, curation, and synthesis. That is a different job, potentially a more interesting one, but it requires a period of unlearning.
Paul Beswick at Marsh understood this instinctively. With 95,000 people in the organisation, he did not mandate AI use or frame it as transformation. He simply made it accessible, surrounded it with clear responsible-use guidelines, and kept the stakes low. Over 300 employee-built AI applications emerged in the first year, with more than one million hours saved.
Ally Bank took a different approach to the same challenge. They went directly after their most sceptical customer service agents, giving them personalised experiences that demonstrated AI’s value on their own terms. Training began not with how the tool works but with what are you afraid of — because until those fears are addressed, no amount of technical demonstration will shift behaviour.
The finding from Ray Eitel-Porter‘s work on AI governance reinforces this: just a few hours of well-structured training moved employee confidence in AI dramatically. The investment required is modest. The obstacle is usually the belief that there is no time — which is itself a reflection of how seriously leadership is taking the transformation.
The Playground Paradox: Why Boundaries Unlock Innovation
Charlene introduced a concept that has since appeared across multiple Arkaro Insights conversations: the Playground Paradox. When you put children on an open playground with no boundary, they cluster in the middle around the familiar equipment. Build a fence, and they explore the entire space — right up to the edges.
The same is true in organisations. Without clear boundaries on AI use, people impose their own imagined limits, which are almost always narrower than the organisation actually intends. The result is timidity dressed up as prudence. AI governance is most often framed around what you cannot do. The more powerful frame is: here is what you can do — all of it, as fast and as far as these guardrails allow. Go.
This is closely related to Charlene’s argument about speed. Safe, as she puts it, is slow. And slow, in a world where AI capabilities are compounding, is a competitive choice — just not a conscious one.
The Data Readiness Myth
One of the most common reasons organisations delay AI adoption is data quality. We cannot start yet because our data is not ready. Charlene’s response is unambiguous: your data will never be ready. It will never be perfect. Waiting for that day is ceding your competitive advantage to those willing to start with messy data.
The more productive question is not whether your data is clean but what is the minimum data quality you need to deliver value against your highest-priority AI application. Identify your most valuable use case. Establish the minimum viable data quality to begin. Start there. The irony is that AI itself is often the most effective tool for cleaning the data you have.
This connects to Charlene’s broader challenge to the readiness assessment and feasibility study as planning instruments. What do you find in a readiness assessment? That you are not ready. What does a feasibility study tell you? That it is not feasible. If the AI application is genuinely important to your business strategy, you will find the resources, build the skills, and make the investments. Feasibility is not the constraint — will is.
Speed Is the New Competitive Moat — and Annual Budgets Are the Enemy
Charlene is direct about the mismatch between AI’s pace of development and the annual planning cycle that still governs most large organisations. The 12-month budget was designed for a world that no longer exists — one where you could plan in autumn for what would happen in December of the following year.
In a single conversation with us, she referenced Claude Cowork and OpenClaw as two developments that had already upended assumptions about agentic AI within a matter of weeks of each other. No annual plan accommodates that pace of change.
Her recommendation is a quarterly approach: a rolling 18-month view reviewed and refreshed every quarter. The strategic objectives stay stable. The roadmap for achieving them is written in pencil, not carved in stone. McKinsey, she notes, is already rejigging its AI roadmap almost daily — not because the destination changes, but because the path does.
For those of us who work with clients on integrated business planning, this resonates deeply. The technology to support quarterly replanning already exists. The obstacle is not capability — it is the willingness to let go of false certainty.
How to Start: The 90-Day Blueprint
For leaders who feel behind, Charlene’s message is reassuring in its precision. You are not too late. And the answer is not to consume more AI content, take more courses, or subscribe to more newsletters. The answer is to identify your biggest business problem and ask how AI can help you address it.
Sales cycles too long? Customer churn too high? Product launch timelines too slow? Start there. In 90 days, you can identify the problem clearly, understand how AI can address it, and establish the building blocks you need to execute. That is the structure of the book — and it is deliberately designed to cut through the overwhelm of AI as a general phenomenon by grounding it in the specific problems that leaders already understand.
AI adoption, Charlene observes, does not vary by age or role. It varies by disposition — specifically, how open you are to change and how comfortable you are with uncertainty. If you are asking the question, your disposition is already right.
The Superhuman Workforce
The concluding theme of the conversation was the concept Charlene and Katia call the superhuman workforce. Not humans replaced by AI, and not humans using AI as a faster typewriter, but a genuine integration of human and machine intelligence where each amplifies the other.
AI handles the computational, repetitive, and administrative load. Humans bring empathy, judgement, creative problem-solving, and wisdom — capabilities that are not just residually human but become more valuable precisely because AI handles everything else. The goal is not to become more like AI. It is to become more deeply human.
IBM’s decision to double down on entry-level hiring — bringing in graduates into coding roles and rewriting those job descriptions to assume AI fluency — is an early signal of what this looks like in practice. The jobs are not the same as they were a year ago. But the jobs exist, and they are more interesting.
This is what we consistently find at Arkaro. The organisations that navigate change most successfully are not those that minimise disruption but those that create the conditions — clear direction, genuine participation, psychological safety, and leadership that models the behaviour it wants to see — where people can step into something new without abandoning what made them valuable in the first place.
Listen to the Full Conversation
Watch and listen to the full episode with Charlene Li on the Arkaro Insights podcast:
📺 YouTube: www.youtube.com/@arkaro
🎧 Audio: arkaroinsights.buzzsprout.com
📖 Get the book: winningwithaibook.com
Related Reading from Arkaro Insights
On AI as a Leadership and People Challenge
- How to Use AI Without It Going Wrong | Ray Eitel-Porter on AI Governance — Charlene Li cites Ray Eitel-Porter’s finding that a few hours of structured training moved employee confidence in AI from 15% to 80-90%. This is the full conversation on responsible AI governance and what organisations need to do to make adoption both safe and fast.
- Rewire or Retire: Why AI Is a Leadership Issue, Not a Technology Problem– Like Charlene, Marco Ryan, Former BP Chief Digital Officer, argues AI isn’t a technology problem—it’s a leadership issue. Leaders must be digitally curious.
- AI & the Octopus Organisation with Stephen Wunker — why most companies are sprinkling AI dust on existing processes rather than transforming their model, and what adaptive organisations look like instead. Essential companion reading to the Charlene Li episode.
- MIT Research Shows What We’ve Been Saying: Why 95% of AI Implementations Fail — the neuroscience evidence base for why AI adoption fails on human factors, not technology. Directly supports Charlene’s central argument.
- Why AI Implementations Fail — and How PEOPLE Can Fix It — the PEOPLE framework for people-centred AI change management, with practical guidance on turning resistance into momentum.
- AI as an Amplifier: Why Organisational Design Determines AI Success — AI amplifies what is already there: effectiveness in aligned teams, dysfunction in unclear ones. Rich Allen on why org design must precede AI deployment.
On the Neuroscience of Change and Psychological Safety
- The Neuroscience of Collaboration: Why Your Brain Still Thinks It’s on the Savannah — Hilary Scarlett on the SPACES model and why human brains in threat mode cannot collaborate, innovate, or engage with change. Directly underpins Charlene’s arguments about identity threat and the fear-based resistance to AI.
- Constraints, Playfulness and Ethics: Three Lessons for Leading in the AI Age — Dr Vlad Glaveanu on why constraints enable rather than block creativity, and why playfulness and psychological safety are non-negotiable in AI-age leadership. A strong complement to the Playground Paradox.
On Adaptive Organisations and Planning
- Why Traditional Management Fails — and How Adaptive Organisations Succeed — the case for moving from control and hierarchy to trust and flow in complex environments. The organisational prerequisite for the quarterly AI planning approach Charlene advocates.
- Make Strategy Work: The Power of Integrated Business Planning — how rolling planning cycles and cross-functional alignment create the conditions for AI to deliver real business value, rather than sit as a separate technology initiative.
- AI and Strategy: Why Implementation Remains Human — AI can generate sophisticated strategic analysis but cannot replace human collaboration in implementation. The bridge between Charlene’s argument about AI-enabled business strategy and the practical work of making strategy stick.
Is Your Organisation Ready to Move from Curiosity to Implementation?
The 90-day window Charlene describes is not a marketing construct — it is the minimum viable timeframe for turning AI intent into business results. Most organisations already have the data, the technology access, and the talent. What they often lack is a clear line from AI capability to business priority, and the change management approach to bring people through the transition.
At Arkaro, we work with leaders in agriculture, food, and chemicals to do precisely that. Our collaborative do-it-with-you approach connects the technical and the human — ensuring that AI investments are grounded in real business problems and that the people responsible for delivering value are genuinely enabled to do so.
We don’t just coach — we get on the pitch with you.
Contact Mark Blackwell: mark@arkaro.com
Connect on LinkedIn: Mark Blackwell
Subscribe to the Arkaro Insights newsletter: Arkaro Insights on LinkedIn
💬 What is your biggest AI implementation challenge right now? Share your experience in the comments.
Do you know three people who would benefit from reading this? Please share it with them. Thank you.
