AI disruption and the collapse of certainty

Modern businesses run on systems of certainty. AI is breaking them. Strategic planning, objectives and key results, and service-level agreements: They are all designed to reduce ambiguity and increase predictability. They let us set goals, track progress and deliver on expectations. But these systems assume that the future will resemble the past closely enough for plans and results to stay aligned. That assumption no longer holds. The AI disruption and rapid evolution have destabilized every business. The ground shifts weekly. Planning systems built for stability cannot stay upright. I’ve been through other major revolutions — cloud, mobile, e-commerce. None of them shattered the strategic playbook like AI. Even companies celebrated for disciplined, input-driven planning have found that what worked for decades falls flat in the AI era. AI is a disruptor, but you can use it as a thinking partner to simulate futures, stress-test ideas and surface blind spots at a scale traditional strategy teams cannot match.

From single-path to multipath planning

Most companies treat scenario planning as a formality: They add a few “what-if” slides to a singular, polished path forward. When Google’s DeepMind team plans, they don’t pick the “most likely” future. They map multiple plausible futures simultaneously and look for moves that work across scenarios. AI makes this practical for any business. Instead of weeks of manual modeling, you can use reasoning models to simulate competitive responses, customer behaviors or macroeconomic shifts in minutes. The point of multipath planning isn’t to predict what will happen; it’s to stop assuming that you know or can know. When the cost of simulating options drops, there’s no excuse for skipping the exercise.

From fixed bets to no-regrets moves

In fast-changing environments, the worst-case scenario isn’t one of being wrong. It’s being locked in when you’re wrong. Traditional planning rewards decisiveness: Pick a direction, get buy-in, execute. In an AI-accelerated economy, that’s a brittle posture. That’s why I’ve started thinking in terms of no-regrets moves: actions that retain value across multiple possible futures. This is how venture capitalists (VCs) operate. They fund a portfolio of startups, each investment a calculated exposure to a potential positive outcome. VCs know that a single win can generate more returns than all the other investments combined. The VC model’s power is optionality in distributed risk, and the ability to capture asymmetric upside when it happens.

From long cycles to short loops

If you’ve worked with software projects, you’ll know Agile development: small sprints, frequent releases, constant communication. We can use the same mindset for almost any business project in the AI era. Shifts in competition, regulation and customer behavior can make a plan stale within weeks. Short loops mean breaking a plan down into smaller, self-contained components, which is a job in itself. It also requires that you build in sensitivity to subtle market signals, replacing an unwarranted reliance on planning with real-time pattern-matching and adaptation.

From top-down to adaptive systems

Traditional planning assumes that direction flows from leaders to doers. In a stable environment, that can work. In an AI-speed environment, that creates lag. Many organizations try to include input from multiple stakeholders through rituals like the “disagree and commit” stage of planning in which teams air concerns, leadership decides and everyone moves forward together. That’s valuable for alignment, but it often happens once, at the start of the cycle. If the ground shifts two months later, commitment turns into inertia. Adaptive systems treat planning as a living, networked capability. Leaders still set priorities and guardrails, but teams adjust tactics in real time based on new signals, which can come from AI at ground level. Amazon has an internal team of economists that illustrates this model. They don’t dictate strategy but constantly read the environment and signal where change is happening. AI can now supplement these skills, or give companies without Amazon’s resources the beginnings of similar capability. To organize this input, you don’t need a chief AI officer at the top. Adaptability comes from AI being embedded in everyone’s work and thinking. It makes the organization harder to surprise, faster to respond and more likely to capitalize on emerging opportunities.

Meet your new high-priced planning consultant

The old way of planning —  reconciling the goals from the C-suite with the truths from the ground — meant suffering spreadsheets, meetings and — worse yet — consultants. Instead, AI can be a powerful consultative presence, an always-on McKinsey, absorbing and processing targets, considering data from the field, and generating and adapting reconciled paths in real time. In an AI-driven economy, the leader’s job is to build an operating model that can sense, interpret and respond to change: Radar, not roadmap. We have to know when to lean on AI for insight, when to question it, how to make it part of the team without letting it think for us. And, critically, it can show us how to teach our teams this discernment. AI is a pattern-finding engine, not an oracle. It can expose possibilities and risks you might miss, but it also inherits the limits and biases of the data it sees. Planning is dead, at least the kind of planning that pretends to control the future with a single path. Long live planning that’s AI-augmented, iterative and deeply human-led.