One of the most common reactions people have when they first hear about multi-agent simulation is that it sounds complicated. Building thousands of AI agents, constructing knowledge graphs, running simulations -- surely that takes weeks of setup and a team of data scientists?
It doesn't. With Foretide, you can go from raw data to a full prediction report in about five minutes. Here is exactly how it works.
Step 1: Upload Your Data
Start by uploading the documents that describe your situation. These can be:
- Strategy documents -- business plans, competitive analyses, market research
- Reports -- quarterly results, industry reports, analyst coverage
- Internal memos -- meeting notes, project briefs, policy documents
- Organizational data -- org charts, stakeholder maps, partnership agreements
You do not need perfectly structured data. Foretide works with the messy, real-world documents that already exist in your organization. PDFs, Word documents, and text files all work.
The key is relevance. Upload the documents that contain the context for the question you want answered. If you are asking about a product launch, include your market research, competitive analysis, and launch plan. If you are asking about an organizational change, include the relevant org charts, policy documents, and stakeholder communications.
Step 2: Ask Your Question
Once your documents are uploaded, type your question in plain language. No query syntax. No configuration files. Just ask what you want to know.
Good questions are specific and outcome-oriented:
- "What will happen to our market share if we raise prices by 15%?"
- "How will employees react to the proposed remote work policy?"
- "Which competitors are most likely to respond aggressively to our market entry?"
- "What is the probability that this merger will face regulatory resistance?"
The more specific your question, the more focused and useful the simulation results will be.
Step 3: Watch the Knowledge Graph Build
After you submit your question, Foretide begins extracting entities and relationships from your documents. You can watch this happen in real time as the platform constructs a knowledge graph that maps out the people, organizations, products, regulations, and events relevant to your scenario.
This step typically takes 30 to 60 seconds depending on the volume of documents. The knowledge graph is the foundation that ensures every simulated agent has access to accurate, contextual information rather than generic assumptions.
Step 4: Generate Agents
Foretide automatically creates thousands of intelligent agents based on the entities and dynamics identified in your knowledge graph. Each agent gets:
- A role and perspective relevant to your scenario
- Knowledge drawn from your specific documents
- Decision-making logic that reflects their position and motivations
- Relationships with other agents that mirror real-world connections
You do not need to configure individual agents. The platform handles this automatically, though advanced users can adjust agent parameters if they want more control.
Step 5: Run the Simulation
With agents generated, the simulation begins. Agents interact with each other, make decisions, respond to events, and influence one another -- all within the context of your scenario. The simulation runs multiple iterations to capture the range of possible outcomes.
This is where the power of multi-agent simulation becomes visible. Instead of calculating a single answer, Foretide explores the space of possibilities, identifying which outcomes are most likely and what conditions lead to each one.
Step 6: Read Your Report
When the simulation completes, you receive a structured report that includes:
- Primary outcomes -- the most likely results with probability ranges
- Key drivers -- the factors that had the greatest influence on outcomes
- Risk scenarios -- less likely but high-impact possibilities to watch for
- Agent insights -- notable behaviors and decision patterns that shaped results
The report is designed to be actionable. It does not just tell you what might happen -- it tells you why, and what you can do to influence the outcome in your favor.
Tips for Best Results
Be generous with context. The more relevant documents you upload, the richer the knowledge graph and the more realistic the agents. A simulation based on three documents will be less nuanced than one based on thirty.
Ask one question at a time. Focused questions produce focused simulations. If you have multiple questions, run separate simulations for each.
Include opposing viewpoints. If you have documents that present different perspectives on your scenario -- bullish and bearish analyses, internal disagreements, competitor materials -- upload them all. Diverse inputs produce more realistic agent populations.
Iterate and refine. Your first simulation gives you initial insights. Use those insights to refine your question or add more context, then run again. Each iteration deepens your understanding.
What Kind of Data Works Best
Foretide works with any text-based documents, but some types are particularly valuable:
- Documents that describe relationships between stakeholders
- Materials that reveal motivations and incentives of key actors
- Analysis that captures market dynamics and competitive positioning
- Historical records that show how similar situations played out before
You do not need quantitative datasets or structured databases. Foretide's strength is extracting intelligence from the qualitative, narrative documents that contain the richest context about how your world actually works.
Ready to Try It?
The fastest way to understand what Foretide can do is to experience it yourself. Sign up for the waitlist and you will be running your first simulation in minutes.
