> For the complete documentation index, see [llms.txt](https://docs.monnfts.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.monnfts.com/mirror-protocol/how-mirror-simulates-the-future.md). # How Mirror Simulates the Future Mirror does not attempt to predict the future by generating one answer from one prompt. Instead, it builds a structured representation of a decision environment, identifies the actors and forces that matter, simulates how they may react over time, and turns those interactions into an inspectable decision forecast. The basic principle is simple: **Model the environment.**\ **Model the actors.**\ **Simulate the reactions.**\ **Inspect what emerges.** This allows teams to explore not only what might happen, but also **why it might happen, who could influence the outcome, and which assumptions could change the result.** ### Mirror Simulates by Making the Hidden Scenario Structure Visible Most strategic decisions contain an invisible system underneath them. Consider a product launch. The outcome may depend on: * Customer expectations * Price sensitivity * Competitor behavior * Market timing * Brand trust * Influencers * Distribution * Product positioning * Media coverage * Existing alternatives A normal AI prompt often compresses all of those factors into a single response. Mirror takes a different approach. It attempts to make the underlying decision environment explicit before interpreting the outcome. The objective is to expose: **Actors → Incentives → Relationships → Reactions → Outcomes** Once that structure becomes visible, the scenario can be simulated rather than simply described. ## An AI Scenario Simulation Turns Evidence Into a Reviewable Model of Reaction At a high level, a Mirror simulation can be understood through four components. | Component | What it represents | | ------------ | ------------------------------------------------------------------------------------------------------ | | **Input** | Evidence, assumptions, decision context, market information and the action being tested | | **Actors** | Customers, competitors, stakeholders, communities or other entities capable of influencing the outcome | | **Dynamics** | How actors interact, respond, adapt and influence one another | | **Output** | Plausible reaction paths, risks, assumptions, opportunities and decision-relevant insights | The important difference is the middle. Mirror does not simply transform: **Input → Answer** It explores: **Input → Actors → Interactions → Emerging Outcomes** That interaction layer is where simulation becomes useful. ## From Real-World Evidence to Simulated Reaction A Mirror simulation can be broken into five core stages. ### Step 1: Define the Scenario Every simulation starts with a decision. The user defines the action or change that should be tested. Examples include: > What happens if we increase our subscription price by 20%? > How might customers react if we launch this product in Dubai? > What happens if our main competitor releases a cheaper alternative? > How could different stakeholders respond to this announcement? The objective is not to describe an entire industry. It is to establish a **bounded decision environment**. A good scenario identifies: * The proposed action * Relevant audience * Market or environment * Important constraints * Time horizon * Decision objective This becomes the foundation of the simulation. ## Step 2: Identify Actors, Motives and Constraints Once the scenario is defined, Mirror identifies the actors capable of influencing the outcome. Depending on the simulation, these could include: * Existing customers * Prospective customers * Competitors * Investors * Employees * Partners * Regulators * Journalists * Influencers * Communities * Distribution channels * Market segments But identifying actors alone is not enough. Each actor exists within a set of incentives and constraints. A customer may care about: * Price * Trust * Convenience * Status * Switching costs * Product quality A competitor may care about: * Market share * Customer retention * Pricing pressure * Brand positioning * Strategic differentiation An investor may care about: * Growth * Risk * Margins * Market size * Execution capability Mirror attempts to model these differences because different incentives produce different reactions. ## Step 3: Build the Scenario Model The next stage connects the pieces. Mirror structures relationships between: * Actors * Claims * Incentives * Market forces * Evidence * Constraints * Dependencies * Potential reactions Think of this as a **scenario graph** or decision model. It makes relationships that may otherwise remain hidden easier to inspect. For example: **Price increase** ↓ Price-sensitive customers become more likely to churn ↓ Competitor gains an acquisition opportunity ↓ Competitor increases advertising ↓ Customer switching becomes easier ↓ Retention risk increases The important insight is not merely that a price increase could create churn. It is the **reaction chain** created after the original action. This is where simulation begins to reveal second-order effects. ## Step 4: Run Multi-Agent Interaction After the decision environment has been structured, Mirror can simulate how the actors respond. This is the multi-agent layer. Rather than asking one AI model to represent every perspective simultaneously, different agents can represent different participants or behavioral perspectives within the scenario. They react to: * The original decision * Available evidence * Their own incentives * Environmental conditions * Other actors * Previous reactions The simulation can continue through multiple rounds. For example: #### Round 1 A company increases its price. Customers react. #### Round 2 Competitors observe customer dissatisfaction and introduce promotions. #### Round 3 Customers compare alternatives. #### Round 4 The company changes messaging or offers incentives. #### Round 5 Different customer segments respond differently. The outcome is no longer generated from one isolated action. It emerges from interaction. This is why multi-agent simulation is particularly useful for situations where: **reaction changes the outcome.** ## Step 5: Convert Simulation Into a Decision Forecast The raw interaction between AI agents is not the final product. The objective is to convert simulated behavior into something a decision-maker can use. Mirror analyzes the simulation and surfaces patterns such as: * Most plausible reaction paths * Important disagreements * Adoption barriers * Competitive threats * Behavioral differences * Reputation risks * Price sensitivity * Trust signals * Opportunities * Critical assumptions * Missing evidence * High-impact uncertainties The resulting forecast is better understood as a **decision map** rather than a single prediction. It should help answer: **What might happen?** **Why might it happen?** **Who could drive the outcome?** **What could change the outcome?** **What should we test next?** ## The Scenario Model Is an Audit Surface One important advantage of structured simulation is inspectability. When an AI system provides only a final recommendation, it can be difficult to understand what produced the answer. A simulation should expose more of the decision structure. Before trusting the result, operators should be able to examine: * Which actors were included * Which actors were excluded * What relationships were modeled * Which incentives were important * What assumptions influenced the run * Which evidence supported those assumptions This makes the scenario model an **audit surface**. Instead of simply asking: > Do I believe the AI? the operator can ask: > Does the simulated world represent the real decision environment well enough? That is a much more useful question. ## Agent Rounds Expose Movement Static analysis is useful for understanding a situation at one point in time. Simulation adds another dimension: **movement.** Actors react. Other actors observe those reactions. Behavior changes. Information spreads. Competitors adapt. Trust increases or decreases. Narratives evolve. This creates a distinction between: **State analysis** and **dynamic simulation.** A normal market report may tell you that customers are price sensitive. A simulation can explore what happens after: 1. You increase the price. 2. Customers complain. 3. A competitor sees the opportunity. 4. The competitor lowers its price. 5. Online discussion amplifies the difference. 6. Your premium customers remain. 7. Price-sensitive customers begin switching. The sequence matters. Sometimes the most important outcome occurs several reactions after the initial decision. ## Why Multi-Round Simulation Matters Many strategic effects are second-order or third-order. The immediate reaction may not be the most important one. Consider a company launching an aggressive discount. #### First-order effect Customers respond positively. #### Second-order effect Competitors lower their prices. #### Third-order effect The entire category becomes more price-sensitive. #### Fourth-order effect Margins decline across the market. A single AI answer might recognize some of these possibilities. Multi-round simulation is specifically designed to explore how one reaction can create another. This allows Mirror to investigate **reaction paths**, not only direct consequences. ## A Useful Forecast Answers Bounded Questions A Mirror forecast should not attempt to answer everything. The strongest simulations focus on questions that can be connected to an actual decision. For example: #### What changed because of the decision? Which actors altered their behavior after the simulated action? #### Who reacted most strongly? Which customer groups, competitors or stakeholders were most sensitive? #### Which risks repeated across different paths? A risk that appears across multiple simulations may deserve more attention. #### Which assumptions influenced the outcome? What must be true for the simulated scenario to occur? #### Where does uncertainty remain highest? Which parts of the decision require more evidence? #### What should we test next? The simulation should lead toward another decision, experiment or evidence-gathering activity. ## Mirror Is Not a Single-Future Prediction Engine Complex systems rarely have one inevitable future. Market outcomes depend on: * Behavior * Timing * Competition * Information * External events * Incentives * Randomness * Decisions made by other participants For this reason, Mirror should not attempt to produce: > The future will be X. A better framing is: > Given these assumptions, actors and conditions, these are the reaction paths worth considering. That distinction matters. Simulation is most valuable when uncertainty remains visible. ## Scenario A, B and C Can Produce Different Futures Consider a company planning a new software product. The baseline offer is: **$99/month** Mirror could simulate multiple environments. #### Scenario A — Baseline Launch at $99. #### Scenario B — Lower Price Launch at $69. #### Scenario C — Premium Strategy Launch at $129 with stronger premium positioning. #### Scenario D — Competitive Response Launch at $99 while the largest competitor reduces its price by 25%. #### Scenario E — Stronger Social Proof Launch at $99 with significantly stronger customer evidence and testimonials. The purpose is not necessarily to identify one mathematically perfect answer. The objective is to understand: **Which variables materially change behavior?** That information can be extremely valuable before a real launch. ## Operator Checks Before Trusting a Simulation Human judgment remains an essential part of the workflow. Before making a decision based on a simulation, review the environment that produced it. ### Are the right actors represented? Missing an important stakeholder can distort the reaction path. ### Are actor incentives realistic? Agents should reflect plausible motivations rather than arbitrary personalities. ### Is the evidence relevant? More data does not necessarily mean better simulation. Relevant evidence matters more. ### Are assumptions visible? Hidden assumptions make forecasts difficult to evaluate. ### Are alternative scenarios considered? One simulation should not automatically become the decision. Test the variables that matter. ### Are uncertainty and evidence gaps visible? A strong decision system should expose uncertainty rather than disguise it. ## What Makes a Simulated Forecast Different From a Summary? A summary describes existing information. A simulation asks what may happen **after something changes**. For example, a summary might say: > Customers in this category are price sensitive and competitors offer cheaper alternatives. A simulation asks: > What could happen if this company increases its price by 20% while competitors maintain their current prices? The difference is: **Summary → What is known** **Simulation → What could emerge** Both are useful. They solve different problems. ## Simulation Is Not the Same as Prediction Prediction generally tries to estimate a future outcome. Simulation explores the system that could produce different outcomes. That means Mirror is interested not only in: **the result** but also: **the path to the result.** For strategic decision-making, that path can be as valuable as the final forecast. Knowing that demand might fall is useful. Understanding that demand might fall because price-sensitive customers move first, competitors amplify the price difference, and social proof deteriorates gives a team significantly more information about what it could do next. ## When Should You Run Another Simulation? Run another scenario when you discover a variable capable of materially changing the outcome. Examples include: * Different price * Different launch timing * Different positioning * Different customer segment * New competitor * New market condition * Different message * Stronger evidence * Regulatory change * Different distribution strategy A useful process might look like: **Simulation 1** ↓ Discover critical assumption ↓ **Simulation 2** Change that assumption ↓ Discover sensitive actor ↓ **Simulation 3** Change that actor's environment ↓ Compare outcomes This turns simulation into iterative decision exploration. ## Questions About the Mirror Simulation Engine ### Is Mirror predicting the future with certainty? No. Mirror generates plausible reaction paths based on the scenario, evidence, actors and assumptions available to the system. Its purpose is to help decision-makers understand possible outcomes and uncertainty—not claim certainty about future events. ### What makes Mirror different from a normal AI analysis? Traditional AI analysis generally evaluates information and generates an answer. Mirror creates a structured scenario, models relevant actors and incentives, and explores how reactions may evolve through interaction. This allows teams to investigate dynamic behavior rather than only static information. ### Why does the scenario structure matter? Because the quality of a simulation depends on what the system believes exists inside the simulated world. If important actors, incentives, constraints or relationships are missing, important reaction paths may also be missing. Making the structure inspectable allows operators to challenge the simulation before relying on its output. ### Why do multiple agent rounds matter? Many outcomes emerge through chains of reaction. Customers respond to companies. Competitors respond to customers. Media responds to controversy. Markets respond to competitors. Multiple simulation rounds help reveal these second-order and higher-order effects. ### Does Mirror replace traditional forecasting models? No. Statistical forecasting remains highly valuable for problems with strong historical datasets and measurable patterns. Mirror is designed for a complementary class of problems where **behavior, interaction and strategic reaction influence the outcome.** ### When should I run another simulation? Run another simulation whenever an important assumption, variable or actor could materially change the outcome. Comparing scenarios is generally more useful than assuming a single simulation represents the future. ## From AI Answers to Simulated Worlds Generative AI made it possible to ask increasingly sophisticated questions. Simulation represents the next step. Instead of only asking an AI system: > What should we do? teams can create a representation of the environment and ask: > What might happen if we do this? Then change the conditions. Run it again. Compare the outcomes. Challenge the assumptions. Gather additional evidence. Repeat. This transforms AI from a system that produces answers into an environment for **testing decisions**. ## Simulate the Reaction Before Making the Decision Every strategic decision eventually enters a system filled with other actors. Customers respond. Competitors adapt. Investors reassess. Communities discuss. Markets move. Narratives change. The decision itself is only the beginning. Mirror is designed to explore what may happen **after the decision enters that system**. **Model the environment.** **Simulate the actors.** **Observe the reactions.** **Inspect the assumptions.** **Compare the possible paths.** Then make the decision with a better understanding of what could happen next. **Don't just predict the future. Simulate how it could unfold.**