A strategy that thinks.

Noostrategy is an agentic strategy practice. It makes live strategies, with humans and agents, and keeps them in play. Strategy becomes the organization's shared thinking layer, guiding the work and learning from it. We write about agentic strategy, we have built the instruments for it, and we run it with organizations.

Why noo?

Noo comes from nous, the Greek word for mind. A century ago Vladimir Vernadsky described the planet in three layers: the geosphere of rock, the biosphere of life and, emerging from them, the noosphere of thought. Once thought appears, it starts reshaping the other two.

Noostrategy is that layer in an organization: its mind, created by humans and agents, shaping what the organization is and what it does.

A live strategy is how it is formed and how it keeps thinking: guiding the work, learning from it, and never finished, because the world it thinks about is not finished either.

Vladimir Vernadsky, The biosphere and the noosphere (American Scientist, 1945).

What has strategy lost?

Strategy has lost its agency. To act on the world with intention, a strategy needs to rest on a difference, something of your own, and to stay open to the response, so it can react when the world answers back. Lose either and the strategy has lost its agency, and stopped acting without noticing.

Is it different or indifferent? A live strategy is different because it rests on a leap somebody made, a bet on why this organization and not another. The humans keep it different, working with the agents: they want something, bet on it and make the call. Automating the strategy makes it indifferent, a default nobody chose, and the models have made the default free: ask one for a strategy and you have it in seconds, and so does everyone else in the category, because wanting is not in the data. That is our first enemy, autostrategy, the plan on autocomplete.

Is it open or closed? The agents keep a live strategy open: they read everything, keep the maps current and bring the world back in every day, which no team has the hours to do. Most strategy has been made closed: thought once at an offsite, printed, carried through the year on habit. That is our second enemy, necrostrategy, the plan stuck in the past. It made sense when thinking was scarce and the world held still long enough for a plan to last.

Lose both difference and openness and you have the third enemy, zombie strategy: the old deck automated, running faster than anyone can find the off switch. A lot of organizations are about to build one.

Two ways a strategy loses its agency, and one way it keeps it: open, and your own. OPEN CLOSED DIFFERENT INDIFFERENT live strategy your own theory, played by humans and agents autostrategy the plan on autocomplete, refreshed nightly zombie strategy a photocopied plan, on cruise control necrostrategy your own deck, stuck in the past revised by the world sealed from the world a leap somebody made the default nobody chose
Two ways a strategy loses its agency, and one way it keeps it: open, and your own.

What is it built on?

Two bodies of thought sit under the practice, one under each instrument. The latest strategy research, systems research and design research, with complexity science behind them, give Live Strategy its spine: what a strategy is, how change happens, and why a strategy cannot be planned once and executed. Computer science and cognitive science give the Strategy Ensemble its model: how a new idea is formed, and which kinds of thinking a machine can do and which it cannot.

Four ideas come from them, and the practice holds the four together: a strategy is a theory, change is a transition made in three creative acts, only one of the three kinds of reasoning makes anything new, and a strategy is played rather than planned. Each is drawn below with the thinking behind it, and each is worked out in a memo.

What is a strategy?

A strategy is the organization's own explanation of how it will create value in a world that does not exist yet, written so it can be wrong. A plan says what to do. A theory says why it should work, and so it says what would have to be true, which can be checked. So a strategy is a hypothesis, and strategy work is finding the explanation worth betting on, saying what would have to be true for it to hold, and putting it under test.

A theory derived from the data everyone has is the theory everyone else has. To give a lasting advantage it has to reach for a difference, which is why it is chosen from rival theories rather than read off the evidence, with the runner-up kept beside it as a challenger.

This is the theory-based view of strategy, set out by Teppo Felin and Todd Zenger in The theory-based view: economic actors as theorists (Strategy Science, 2017) and in Zenger's What is the theory of your firm? (Harvard Business Review, 2013). The question what would have to be true is Roger Martin's, from the possibilities-based approach in Bringing science to the art of strategy, with A. G. Lafley, Jan Rivkin and Nicolaj Siggelkow (Harvard Business Review, 2012), and in Playing to Win (2013).

Induction reads above the horizon. Abduction proposes what lies below it. a new premise the explanation worth betting on a trend an anomaly a narrative a complaint what the data shows what would explain it
Above the horizon, what the data shows. Below it, what would explain the data: a new premise, the explanation worth betting on.

How is a strategy made?

A strategy is made in three creative leaps. Change in an organization is a transition from one system to another. The present system is how it makes value now. The future system is the one it chooses to grow, already visible in pockets of what it does. The present system keeps itself going: new work tends to get pulled back into the way things are done now. So every bet made in between either builds the future or feeds the present, and from inside it is hard to tell which. That is why the transition is mapped as a system of its own.

Making a strategy is three leaps across those systems, with exploration in between. The problem worth solving: framed from the present as mapped. The theory of success: chosen from the future as mapped. The bets: placed on the transition as mapped. Each is creative rather than analytic: it proposes something that does not exist yet and says how it would hold together. A person makes or picks each one. The ensemble explores the ground between them.

The three systems are Bill Sharpe's Three Horizons, rewritten as a theory to test: Three Horizons: The Patterning of Hope (2013) and, with Anthony Hodgson, Graham Leicester, Andrew Lyon and Ioan Fazey, Three horizons: a pathways practice for transformation (Ecology and Society, 2016). The creative acts are system-shifting design, as Cat Drew, Cassie Robinson and Jennie Winhall described it in System-shifting design: an emerging practice explored (Design Council and The Point People, 2021), with Kees Dorst's Frame Innovation (2015) behind the first of them.

The stroke of a strategy: three systems mapped, three leaps made or picked by a person, the ground between them explored by the ensemble. the present mapped the problem the future mapped the theory the transition mapped the bets a person leaps the ensemble explores a person leaps the ensemble explores a person leaps
The stroke of a strategy: three systems mapped, three leaps made or picked by a person, the ground between them explored by the ensemble.

Where does the machine stop?

The machine stops at the leap. Strategy uses all three of the kinds of reasoning Charles Peirce named. Induction finds patterns in what is observed. Deduction draws the consequences of premises. Both work inside a frame. Abduction changes the frame: it is the leap to a new premise, an explanation worth betting on, and Peirce called it the only logical operation that introduces any new idea. If a strategy is a theory, then making one is an act of abduction, and the other two kinds are its preparation and its test.

Language models are strong at induction and increasingly good at deduction, at a volume no team can match. Abduction splits in two. Generating candidate explanations is something a model does well. Selecting one is not, for two reasons. A model has no stakes: it does not live with the choice. And a model's choice is the likeliest one, which is the choice everyone's model would make, so it cannot be the difference a theory needs. So the agents cover the ground and draft the candidates, and the leap itself, and the choice of it, stay with humans.

Peirce drew the distinction in Deduction, induction, and hypothesis (1878) and named abduction in his Harvard lectures on pragmatism (1903). Where a model stops is the finding of the latest computer science: Tom Zahavy's LLMs can't jump (Google DeepMind, 2026), on the premise no text yet contains. Teppo Felin and Matthias Holweg reach the same place from the strategy side in Theory is all you need (Strategy Science, 2024): prediction from data is not the forward-looking, theory-based reasoning that strategy runs on.

Induction and deduction work inside a frame. Abduction changes it. ABDUCTION the leap INDUCTION patterns DEDUCTION consequences strategy anomalies premises inside the frame
Induction and deduction work inside a frame. Abduction changes it. Visual format after @chthonic__youth.

Why is it never finished?

A live strategy is never finished because the world it is about does not hold still. It runs as a loop. Sensemaking reframes the theory. The theory derives hypotheses. Experiments test them. The evidence comes back as learning, and anomalies travel up to the theory, which is how it stays under test. The loop runs in sessions, by the organization, the strategist and the agents together, and it has no last session.

It has to be a loop, because an organization and its market are complex systems: the effect of a move can only be learned by making it, so a strategy cannot be derived in advance, only played. The shortcut runs from sensemaking straight to hypotheses, from what the data shows to what to do, with no theory in between. It is what a plan does, and it is what a model does when left alone. The loop is what organizations could not sustain by hand, and what the agents can keep running.

The loop is the theory-based view's own circulation of theory, hypotheses and evidence (Felin and Zenger, 2017), worked out as a method by Timo Ehrig and Jens Schmidt in Theory-based learning and experimentation (Strategic Management Journal, 2022). The complexity point is the one David Snowden and Mary Boone made for leaders in A leader's framework for decision making (Harvard Business Review, 2007): in a complex system cause and effect are clear only in hindsight, so the way forward is to probe, sense and respond.

The loop of strategy work, and the shortcut it is tempted by. SENSEMAKING induction THEORY abduction HYPOTHESES deduction EXPERIMENTS evidence reframe derive test learn the shortcut anomalies travel up
The loop of strategy work, and the shortcut it is tempted by. Visual format after @chthonic__youth.

The words we use.

Agentic strategy
Strategy made and played by humans and AI agents together. Not advice on adopting AI.
Live strategy
Open, and your own: a strategy that stays in play, revised on what the market does and what the organization learns, for as long as it is played.
Playing the strategy
The sessions in which the team, the strategist and the agents review what the experiments found, make the calls and write it back. Played as in music, not as in a game.
Necrostrategy
The closed plan: your own deck, stuck in the past. Thought once, sealed from the world, carried on habit.
Autostrategy
The indifferent plan: the plan on autocomplete, refreshed nightly. Open to the world, and nobody chose a word of it.
Zombie strategy
Closed and indifferent: a photocopied plan on cruise control, running faster than anyone can find the off switch.
Theory of success
The organization's own explanation of how it will create value, written so it can be wrong.
Bet
A committed piece of work that moves the organization toward the future system and tests one of the theory's claims.
Experiment
A small test that informs a bet: quick, cheap and easy to stop.
The Strategy Ensemble
The humans and the agents who play the strategy, every role filled twice.
The Strategy Canvas
Where the strategy is made and played, on the organization's own files.
The Strategy Roast
A free web tool at strategyroast.com: eight questions about an organization, answered in a browser, and a written read of where it stands.

Read.

Got a question your current plan can't answer?

Let's play it live. Drop Ville Tikka a line about what is on your mind, or just book a chat from the calendar.