Strategic foresight, explained
What Is Strategic Foresight?
A field guide to seeing structural change early enough to do something about it.
By Daniel Zivkovic, founder of Magma Inc.
Strategic foresight is the disciplined practice of detecting the early, faint signs of structural change while they are still faint, and acting on them while there is still time and room to act cheaply. It is not prediction. It is not a crystal ball, a trend report, or a longer planning deck. It is a way of working that treats the future as something you can prepare for instead of something that happens to you.
The short version. The problem is almost never that the change was invisible. It is that by the time it was undeniable, it was already too late and too expensive to do anything but react. Foresight is the discipline of closing that gap. The tools to do it are now everywhere; the scarce part is the judgment to read a signal early and the nerve to act, which is what used to make foresight a luxury only big institutions could afford, and what is now becoming accessible to everyone else.
The break in the curve
Most planning quietly assumes that next year looks like this year with a little more of everything. Draw the trend line, extend it to the right, allocate the budget. That works right up until it does not, and the moment it stops working is the moment that matters most.
The strategist Igor Ansoff gave this moment a name in 1975: the strategic surprise. A sudden, urgent, unfamiliar change that threatens a major reversal of profit, or offers a large new opportunity, and that arrives faster than the organization can respond to it. The defining feature of a strategic surprise is not that it was unforeseeable. It is that the organization meets it at the moment of truth, when the only options left are expensive ones.
The future, in other words, is not a straight-line extrapolation of the past. Past performance does not predict structural shifts, because structural shifts are precisely the places where the curve breaks. Conventional planning fits a smooth line to history and, by its nature, misses the break.
Weak signals, and the price of waiting for certainty
Many structural changes send out a faint, early trace before they fully arrive. Ansoff called these traces weak signals: vague, imperfect, content-poor early indicators. A piece of background chatter. An isolated anomaly. A niche innovation that looks like it is not your problem. Over time, through an amplification process, the weak signal becomes a strong signal: content-rich, statistically obvious, impossible to ignore. By then it is also a concrete crisis, and you are out of room.
Concrete examples are rarely dramatic at first: a niche regulatory proposal, a competitor hiring for an unfamiliar capability, a supplier changing terms in one region, or customers inventing a workaround because the category no longer fits how they work.
Not every change behaves this way. Some are genuine bolts from the blue, with no forewarning to read. But far more arrive with a quiet trace that went unread than arrive with none at all, and the discipline is built for the larger category, not the rare one.
Ansoff laid this out as a progression of knowledge. At the earliest state, you sense only that something is coming, with no shape and no source. At the latest state, the threat or opportunity is fully concrete and you can calculate its exact impact. The trap is that the two move in opposite directions: the more certain you become, the less time and flexibility you have left. Wait for the certainty that conventional planning wants, and you guarantee that you respond too late, at the highest possible cost. Act on the vague early signal, and your response is unfocused but cheap, and you still have options.
This is the part that does not fit on a dashboard. Your dashboards are built to report strong signals, the things that have already happened. By the time a structural shift shows up clearly in your numbers, it has already taken the cheap, early response off the table. The expensive surprises are rarely the ones nobody could see. They are the ones that sat in the weak-signal band, unread, until they were certain.
Isn't this just hindsight dressed up as foresight?
A fair and important objection, and the field has argued about it for forty years. It deserves a straight answer rather than a sales pitch.
The sharpest version came from the critic Ashley, who argued in 1989 that there is no genuine warning phenomenon at all. People look back after an event, want to find warnings, and retroactively assign meaning to information that was meaningless at the time. The early signal, on this view, is manufactured by hindsight. Others piled on: Makridakis and Heau called weak-signal theory merely an academic idea; Webb asked, reasonably, whether anyone had proven weak signals exist before building methods on top of them. And the practitioners who grant that signals exist point out the real difficulty: the hard part was never detection, it was interpretation, telling a real symptom from arbitrary noise, and convincing leadership to believe the analysis.
These critiques are not a threat to foresight. They are the specification for doing it honestly. If most foresight really is hindsight in a nicer font, then the entire value of the discipline lives in the few things that separate it from hindsight.
What separates a disciplined scan from a story told after the fact:
It scans before the fact, on the record
A signal logged and reasoned about today, with a date on it, cannot be a story you told yourself after the outcome was known. The discipline is in the timestamp.
It is a repeatable method, not a vibe
The legitimate criticism of early foresight was that it was unsystematic, one analyst's intuition, unportable from one situation to the next. A method that runs the same lenses over any domain, every time, is the answer to that charge.
It separates the signal from the wish
Naming false positives as false positives, and holding the line that not every strange new thing matters, is what keeps foresight from collapsing into pattern-matching on noise.
So the honest claim is not that anyone can see the future. The claim is narrower, and it is one you can defend in a boardroom: a disciplined scan, run early and written down with a date on it, gives leaders a real chance to notice weak evidence, tell signal from wish, and keep their options open before certainty arrives and forecloses them. It does not abolish uncertainty. It buys back time and room to move inside it, which is the only advantage that was ever available this early.
Why this is not a software problem
It is tempting, in 2026, to assume this is now a tooling question. Point enough AI at enough data and the signals surface themselves.
Half of that is true. In a 2025 survey of 167 foresight experts across 55 countries, the OECD and the World Economic Forum found that two-thirds already use AI in their work, mostly to do exactly this: scan horizons, cluster trends, and draft scenarios. The tooling has arrived, and it is spreading fast.
But read what those same experts concluded about it. Their verdict, in the report's own words, is that AI is a supplement to human work, not a replacement: it handles the data processing and the first drafts, which frees the expert for the higher-level analysis, interpretation, and critical thinking. They also name what AI is worst at, and it is the part that matters most in foresight: the genuinely novel, low-probability signal that has no precedent in the data it was trained on. AI is the signal-processing exoskeleton. It is not the foresight brain.
So the real barrier was never technical. One of the challenges those practitioners most often report is aligning AI output with human-centred judgment, not running the tool. The demand for one hundred percent certainty before anyone will allocate a dollar, and the quiet resistance to acting on an unfamiliar signal, are human problems, and no subscription removes them. AI can hand every company on earth the same firehose of weak signals. It cannot tell you which faint blip is the iceberg, and it cannot walk into a leadership meeting and make the case for acting before the proof arrives. That discrimination, and that nerve, are the discipline. They are what you are actually buying when you buy foresight.
This is why we will just have someone run it through a chatbot is not the shortcut it appears to be. The chatbot gives everyone the radar. Reading the radar early enough to still have cheap options, and getting the organization to move on what it reads, is a trained practice. The tool is the easy part. It always was.
Democratizing strategic foresight
Now the half that genuinely changed.
Strategic foresight is not new, and it is not unproven in practice. Governments, intelligence services, and a handful of very large corporations have run formal foresight units for decades, because they could afford to. Foresight was a luxury good. It required a standing department of analysts, a research budget, and the patience to fund a function whose payoff is a surprise that did not happen to you. A five-hundred-person company that needed it most could never justify the overhead, so it went without, and absorbed the surprises.
That cost structure is the part that has fundamentally changed. The same AI that hands everyone the radar also takes the laborious half of foresight, the wide scanning and the first-pass synthesis, and makes it something a single senior practitioner can run, rather than a standing team. The scarce, expensive half (the disciplined human judgment about what actually matters) becomes the main thing you pay for, because the labor underneath it has largely been automated. The result is that senior-grade foresight, whose textbook case is Shell's scenario planning in the 1970s and which has mostly lived inside governments and very large corporations, comes within reach of a mid-market firm.
That is the shift worth naming plainly: strategic foresight is being democratized. This is not only our framing. The same OECD and World Economic Forum report describes AI's potential to 'democratize access to anticipatory tools that were once resource-intensive and institutionally bound.' The capability was always real and always valuable; it was simply priced as a luxury, and that price wall is coming down.
Practitioners have a name for the broader version of this idea: futures literacy, set out in UNESCO's Transforming the Future: Anticipation in the 21st Century (edited by Riel Miller, 2018), which holds that the capacity to use the future should belong to everyone who must make decisions under uncertainty, not only the institutions large enough to staff for it.
Where it starts: scan the horizon
You cannot mount a focused response to a threat you cannot yet see. So the discipline does not start with strategy, scenarios, or a war room. It starts one step earlier, with horizon scanning: the standing radar that catches the weak signal while it is still weak, while a cheap, early, flexible response is still on the table.
Horizon scanning is deliberately the first move, not because it is the smallest, but because it is where the leverage is. Every expensive thing downstream (the scramble, the crash response, the strategic call made under a deadline) is the price of having scanned too late. Get the sensing right and early, and everything after it gets cheaper. Get it wrong, and no amount of downstream cleverness buys back the time you lost.
This is the part of the discipline Magma exists to make accessible. We run a structured, repeatable horizon scan on your market, with senior judgment supported by AI-assisted scanning and synthesis, so that a mid-market team can have the early-warning capability that used to require a foresight department. Not a prediction. Not a dashboard of things that already happened. An honest, dated, defensible read of what is forming on your horizon, while you can still do something about it.
See the change forming offshore, and act while it is still cheap to act.
That is all foresight is, and it is enough. Start with a structured horizon scan: a dated, defensible read of what is forming around your market before it shows up in your numbers. The Strategic Foresight offer explains how Magma runs that scan when you are ready.
See the Strategic Foresight offerA field guide, not a sales page.
Sources and lineage
This explainer is grounded in the primary literature of the discipline, including its critics.
Igor Ansoff (1975), Managing Strategic Surprise by Response to Weak Signals, California Management Review. The seminal paper: strategic surprise, the planning paradox, weak signals, the states of knowledge, and Strategic Issue Management.
The critique literature, taken seriously. Ashley (1989) on hindsight bias; Makridakis and Heau (1987) and Webb (1987) on the lack of empirical grounding; Betts (1982) on interpretation as the true bottleneck; later work on signal-versus-noise and false positives. Named here because a discipline is only as credible as its answer to its own best critics.
OECD and World Economic Forum (2025), AI in Strategic Foresight: Reshaping Anticipatory Governance. A joint white paper surveying 167 foresight experts across 55 countries: two-thirds already use AI, and the consensus is that it supplements rather than replaces human judgment.
OECD (2025), Strategic Foresight Toolkit for Resilient Public Policy. An institutional treatment of foresight as a usable capability rather than an academic curiosity.
UNESCO / Riel Miller, ed. (2018), Transforming the Future: Anticipation in the 21st Century. The open-access foundation of futures literacy and the Discipline of Anticipation: using the future to surface hidden assumptions, not to predict it.
The futures-studies tradition of the futures cone: probable, plausible, and possible futures.
By Daniel Zivkovic, founder of Magma Inc.