Decision Design Arena
Play With AI
Before AI Plays
With You
A new decision-design game turns chatbot interaction into a practice for directing intelligence before asking for answers.
By Sathi Vanigasooriar | learn108.press
The conversation begins with a situation.
A 41-year-old operations director at a mid-sized logistics company opens a chatbot and reads the prompt on the screen. Her company has received an acquisition offer. The founding family is divided. A private equity firm has a ten-day exclusivity window. The question looks simple enough, until she realizes the chatbot is not asking her what the company should do.
It is asking her how the decision should be thought through.
She has not been asked to give AI an answer. She has been asked to give AI a better brief.
There is a game called Two-5-Two: Decision Design Arena, and it does not work the way most AI games work. There is no trivia. There is no guessing. There is no reward for typing faster, sounding smarter, or getting the chatbot to produce a polished answer in seconds.
What it rewards is more important: the ability to look at a complex, unresolved situation and construct, deliberately and explicitly, the cognitive architecture that the situation deserves.
The game is built on the Two-5-Two Decision Design Language, developed by Sathi Vanigasooriar, whose book Decision — Instrument of Joy positions Two-5-Two not as a framework, not as a methodology, but as something more fundamental: a grammar. A grammar for thinking. One that, its creator argues, is substrate-independent — meaning it works whether the intelligence being directed is a human mind, a team of people, an artificial intelligence, or the quantum computing systems beginning to emerge at the frontier of machine capability.
That is a large claim. The chatbot game is, among other things, a way of testing it.
How a Session Works
A situation appears in the chat — written with enough texture to feel real. Not a case study with a tidy resolution waiting in the appendix. A genuine human predicament, with multiple legitimate paths, real stakes, and the kind of uncertainty that makes decisions hard in the first place. The acquisition offer. The reluctant co-founder. The hospital’s AI screening tool that has flagged a candidate the hiring manager believes, with conviction, is exceptional.
Every player receives the same situation at the same moment. No one gets a head start. And no one, crucially, is asked to solve it. The first instruction is simple: pause, read, and stay with the situation long enough to notice what kind of thinking it is asking for.
Then the chatbot begins the draft.
The Two-5-Two language has nine elements, organized across three categories. Two states — Pause and Play — describe the quality of attention being brought to a decision. Five actions — Ask, Absorb, Access, Activate, Attune — describe the cognitive moves available within that attention. And two lenses — the Situation Triangle and the Opportunity Triangle — describe the two territories every decision must navigate: what currently is, and what could be.
The player does not drag cards across a board. The player converses with the system. She tells the chatbot which Two-5-Two elements she wants to bring into the decision, how much emphasis each element deserves, and why this situation requires that configuration rather than another.
That choice matters. If she emphasizes Ask, she is saying the decision still needs better questions. If she emphasizes Absorb, she is saying the situation must be taken in more fully before action begins. If she emphasizes Activate, she is saying the decision needs movement, commitment, and experiments in the world. Every move is an argument.
For the acquisition scenario, the logistics director loads up on Absorb and Situation △. She is saying the decision requires full intake of current reality before anything else moves. She uses only a light Opportunity △. She is not dismissing the future. She is saying the future cannot be responsibly imagined until the present is fully understood. She gives almost no weight to Pause, which will later become the most interesting thing about her design.
The Chat Is the Design Surface
The interface is simple because the work is not visual decoration.
The player is inside a structured chat. The chatbot asks what needs attention. The player answers. The chatbot asks what is missing. The player adjusts. The chatbot challenges the design: why this move, why now, why not the opposite?
This is where the game becomes different from ordinary prompting. Most people approach AI by asking for output. Write this. Summarize that. Compare these options. Give me the answer. The Decision Design Arena asks the player to do something upstream of output. It asks the player to shape the intelligence before the intelligence responds.
The chat becomes the design surface. Not because it displays a diagram, but because it forces a sequence of conscious commitments. What is the situation really asking for? Which cognitive moves belong here? Which ones are being avoided? What has been overused? What has been left out?
That is why the interaction is conversational rather than merely sequential. Real decisions do not unfold in neat steps. Multiple things are true at once. The current situation and the possible opportunity have to be held together, in tension, or the decision collapses into either pure reaction or pure fantasy. The chatbot gives the player a living surface on which to hold that complexity, question it, revise it, and then brief the intelligence that will help work through it.
The Player Is Not the Answer Machine. The Player Is the Director.
And here is where the game does something genuinely unusual.
The conversation does not represent the player’s final decision. It does not even represent the player’s opinion. What it represents — and this is the reframe that changes everything about how the game is understood — is a brief. A set of instructions. A commission.
The player is not the answer machine.
The player is the Director.
Everything the player says in the chat becomes a directive to an intelligence that will now do the actual work of decision design. In the current version of the game, that intelligence is an AI. But the deeper claim is that this kind of brief can work across substrates. It can instruct your own mind by telling your brain which cognitive tools to deploy. It can instruct a team by giving the conversation a designed architecture instead of letting it follow the loudest voice. It can instruct an AI system, not in the thin sense of typing a prompt, but in the deeper sense of commissioning a mode of intelligence.
And it is designed to be forward-compatible with quantum computing systems, which do not think in ordinary step-by-step sequences, but explore multiple paths before resolution is required.
This last point is the one that tends to stop people.
The standard understanding of quantum computing is hardware-focused: faster processors, different physics, problems that classical computers cannot solve in viable time. What is less commonly discussed is that quantum systems require a different grammar of instruction. You cannot simply tell a quantum computer to do step one, then step two, then step three. That is not the structure it is built for.
You have to frame the problem in a way that allows simultaneous exploration of multiple solution paths, with conditions for how and when to collapse toward an answer. That framing — that grammar of instruction — is what Two-5-Two is training players to practice in human language first. Not as a technical exercise. As a cognitive one.
The Mirror Arrives
Back to the logistics director. Her brief has been submitted. The AI receives the situation text, the Two-5-Two elements she emphasized, the reasoning she gave, the sequence of her responses, and the gaps that appeared in the conversation.
It does not receive her intentions. It receives what she actually built in language.
The AI reads the design across five dimensions. Coverage — which elements of the language were brought in and whether their presence or absence reads as intentional. Balance — whether both the Situation and Opportunity territories were genuinely engaged. Coherence — whether the design fits this specific situation or could have been used for almost anything. Depth — how many layers of the situation the design actually reaches. And Resolution — whether the design arrives somewhere, whether there is a quality of commitment in it, or whether it remains open indefinitely.
Then it returns something that is less a score than a mirror.
The headline verdict, in her case, is this: A design of genuine situational intelligence, grounded in reality, but not yet ready to move.
The score is 74 out of 100. Strong on Coverage. Strong on Coherence. The AI reads her Absorb-heavy brief as deliberately fitted to a high-information scenario and credits her for it. But her Resolution score is low. Activate appears too late and too weakly to function as commitment. It reads, the AI notes, as deferred rather than deliberate. And the near absence of Pause — which she thought was a sign of readiness — reads differently. In a situation involving a divided family and a ten-day window, the AI says, the missing Pause is not confidence. It is the one thing the situation most needed and did not receive.
She reads this and sits with it for a moment.
Then she says: that’s exactly how I handle acquisitions in real life.
The Recognition Is the Game
This is the moment the game is designed for.
Not the score. Not the competition. The recognition. The specific, uncomfortable, accurate recognition that the way you brief AI often reveals the way you already think — and that you can now see it clearly enough to decide whether it is serving you.
Most professional development operates on the assumption that knowing more improves thinking. Read the case studies. Learn the frameworks. Accumulate the models. This is not wrong, but it addresses only the content of thinking, not its architecture. The logistics director knows a great deal about acquisitions. That knowledge is not what the chatbot revealed. What it revealed was the shape of her cognitive approach — and the shape had a gap in it that her knowledge had been filling in, invisibly, for years.
The game works because it makes that shape visible in conversation. The player cannot hide behind expertise, confidence, or the speed of a clever prompt. The brief is the brief. The AI reads it without deference. And the feedback is specific enough to be developmental rather than merely evaluative.
Played once, this is interesting. Played over time, across different situations at escalating complexity, by a player willing to experiment with different configurations — trying more Pause, trying more Opportunity △, bringing Activate earlier, asking the chatbot to challenge hidden assumptions — it becomes something closer to a practice. A regular rehearsal of the most transferable skill available: the ability to consciously construct the right thinking for whatever is in front of you, and transmit that thinking to whatever intelligence is available to execute it.
Your own mind. Your team. Your AI. Eventually, your quantum systems.
The interface changes. The skill is the same.
The operations director plays again the following week. Different situation — a hospital, an algorithm, a hiring manager with a conviction the system does not share. This time she gives Pause a real place in the brief. She asks the chatbot to slow the decision before it accelerates. She brings Opportunity △ closer to the center of the conversation.
Her score goes to 81.
The AI’s headline reads: A design that has learned to stop before it moves.
She prints the result and puts it on her desk. Not because it is flattering. Because it is true — and because she made it true, deliberately, by learning to brief intelligence before asking intelligence for answers.
knowing how to think before she thinks.