Purpose North Star
The enduring human, customer, or societal outcome that gives the organization direction across businesses.
Moves slowestAI strategy often begins too late. Before asking where to deploy AI, leaders need to understand whether AI has changed the environment enough to move the strategic North Star.
Before you ask what AI can do, ask what AI has changed.
Organizations are racing to become faster, cheaper, and more productive with AI. Those goals matter, but they contain an assumption: that the destination remains correct.
A company can automate the current journey while customers move to a different one. It can optimize a metric whose importance is declining. It can use advanced technology to pursue yesterday's definition of success.
A North Star is useful because it gives an organization an observable point to steer by. But the metric is not the value itself. It is a proxy for a deeper outcome.
The metric still appears to point toward the customer outcome the organization intends to create.
Keep the purpose. Re-test the proxy.
A durable North Star architecture connects purpose to customer value, then surrounds the metric with drivers, guardrails, and assumptions that can be tested.
The enduring human or customer outcome the organization exists to create. Technology may change how it is delivered without changing why it matters.
Do not change direction because AI is fashionable. Reopen the question when AI changes the relationship between the metric and the value it was supposed to represent.
If agents act on a person's behalf, does the old human activity metric still describe value?
Revisit who actually receives value and who or what now performs the journey.Purpose, customer value, North Star metrics, drivers, and guardrails live at different levels of abstraction. The telescope makes those levels visible without collapsing them into one number.
Moves slowest. The enduring human outcome should survive changes in product, interface, workflow, and technology.
Do not turn purpose into a quarterly KPI.
Illustrative health example. The telescope teaches the architecture, not a universal healthcare metric. Each organization needs evidence connecting its own purpose, value outcome, proxy, drivers, economics, and guardrails.
The right North Star depends on the level of the organization you are trying to orient. A coherent product can often steer by one value event. A diversified enterprise may need a different architecture.
The enduring human, customer, or societal outcome that gives the organization direction across businesses.
Moves slowestA small set of strategic outcomes that align multiple businesses without pretending their value mechanics are identical.
Moves with strategic contextA measurable customer-value event inside a coherent product or value system, connected to controllable drivers and guardrails.
Most instrumentableDo not force one metric across systems that create value in fundamentally different ways.
The discipline is the sequence. First identify what became possible. Then ask what changed outside the technology. Inside SHIFT, locate which assumption and which strategic layer the evidence actually invalidates. Only then decide whether the destination moved and what the organization must redesign.
Reasoning, agents, memory, multimodality, generation, autonomy, and falling costs of intelligence expand what machines can do.
Capabilities create possibilities. They do not create strategy by themselves.
Look outside the model. Did customer behavior move? Did an interface disappear? Did something scarce become abundant? Did power migrate to a new intermediary? Then locate the assumption that change invalidates and the strategic layer where that assumption lives.
A shift becomes strategic when it invalidates an assumption the current strategy depends on.
Decide whether the governing direction should be preserved, broadened, or replaced. The goal is not to change the North Star because AI exists. It is to recognize when converging evidence says the old direction no longer fits.
This is the moment of reorientation, and sometimes the correct answer is to keep the Star.
Once direction is clear, rewire journeys, products, workflows, metrics, decision rights, data, technology, governance, capital, incentives, and operating models around it. Then test whether the new system actually produces the intended outcome.
Reorient first. Rewire second. Then learn from the evidence.
The framework has four steps, but it is not linear. Evidence from the System returns to Signal and updates the assumptions behind the next decision.
The most dangerous AI strategy is not always one that moves too slowly. It can be one that uses AI to move faster toward an objective whose strategic value is eroding.
Are we using AI to reach yesterday's North Star faster?
North Star Debt is the accumulated cost of continuing to optimize yesterday's definition of success after the environment has changed.
It appears when metrics, journeys, operating assumptions, and investment logic remain anchored to a world that is disappearing. The organization can appear to improve while becoming less aligned with where value is moving.
The framework separates a strategic decision from an organizational redesign.
Understand the signal, interpret the shift, question inherited assumptions, and decide whether the Star has moved.
SIGNAL + SHIFT + STARRedesign journeys, work, metrics, technology, governance, decision rights, and operating models around the direction that matters now.
SYSTEMThe discipline is to change at the lowest level that adequately responds to the evidence. Sometimes the right move is to stay. Often it is to redesign the route. Only occasionally is it to move the governing direction itself.
The underlying customer value, economics, and source of differentiation still fit. The signal may deserve experiments or hedges, but not a strategic reorientation.
Which assumption would have to become false for staying the course to stop making sense?
Do we have persistent behavioral and economic evidence that the current model is becoming less fit, or are we reacting to novelty?
Change at the lowest level that adequately responds to the evidence.
Polaris Shift is an independent strategic framework created by Darko Tushev. It helps teams, functions, and organizations test the assumptions beneath strategy and decide whether to Stay, Adjust, or Shift when conditions change.
A North Star is the guiding outcome used to make trade-offs. A North Star metric is a measurable proxy for progress toward that outcome. A metric can need replacement while the underlying purpose and outcome remain useful.
Signal asks what became possible. Shift examines what changed because of it. Star considers what should matter now. System identifies what must be redesigned. Evidence from action feeds the next cycle. The 25-minute discussion guide applies these questions to a decision.
Stay when the guiding outcome and core approach remain credible. Adjust when the outcome still holds but the route, capabilities, or measures need to change. Shift when evidence challenges the guiding outcome itself. Name the scope of the decision before choosing a response.
The framework and its tools are evolving decision aids. They do not provide a validated score or prove that changing direction will improve results. Explore the research agenda and evidence standards, compare public company reference points in Company Atlas, or try a fictional decision rehearsal.
What has AI made possible that was previously too expensive, too slow, or too difficult?
What changed outside the technology because that capability now exists?
Which current metric or objective assumes the old environment will continue?
Where is value moving faster than our strategy is moving?
Are we improving the path, redesigning the path, or reconsidering the destination?
If the North Star moved, what systems would have to move with it?
The purpose of AI strategy is not to find more places to use AI. It is to decide where intelligence changes what is worth doing.