# Watching Agents — Full Documentation for LLMs

> AI-powered prediction agents that turn any question about the future into a living, evolving forecast.

## About

Watching Agents is a platform that deploys autonomous AI research agents to monitor, analyze, and predict outcomes for any forward-looking question. Each agent continuously scans real-world signals, builds structured hypotheses backed by evidence, and provides probabilistic forecasts that update in real time.

The platform serves investors, strategy teams, journalists, consultants, enterprises, startups, and entrepreneurs who need to stay ahead of change.

## Core Concepts

### Agents (Topics)
An agent is an AI-driven research unit assigned to a specific question or topic. Each agent maintains:
- **Overall Probability**: The aggregated likelihood of the predicted outcome (0–100%)
- **Confidence Level**: How much evidence supports the current probability
- **Change Velocity**: How fast the prediction is shifting
- **Trending Status**: Whether the topic is currently seeing significant activity

### Hypotheses
Each agent generates multiple hypotheses — possible future outcomes. Every hypothesis tracks:
- Probability (0–100%)
- Confidence level
- Evidence strength
- Supporting and contradicting evidence
- What would confirm or disprove it
- Historical snapshots showing how the prediction evolved over time

### Evidence
Evidence items are linked to hypotheses and include:
- A claim with source citation
- Stance: supporting, contradicting, or neutral
- Strength and relevance scores
- Source type and attribution

### Drivers
Key factors that influence the outcome. Each driver tracks:
- Current state and trend direction (rising, falling, stable)
- Impact strength on the overall prediction
- Lag time before effects materialize
- Watch signals to monitor

### Watch Signals
Specific indicators the agent monitors for changes. Each signal has:
- Type classification (regulatory, technical, market, political, social, environmental)
- Importance ranking
- Description of what to watch and why it matters
- Linked hypothesis it would impact

### Sources
Tracked information sources with:
- Credibility rating
- Bias notes
- Role in the prediction model
- Last activity timestamp

## Categories

Agents are organized into categories such as:
- Technology, AI, Space, Climate, Geopolitics, Economy, Health, Energy, Defense, Regulation, and more.

## Content Types

### Agent Pages (`/agent/{slug}`)
Each agent page includes tabs for:
- **Overview**: Summary metrics, current state, key hypotheses
- **Hypotheses**: Detailed hypothesis cards with evidence
- **Drivers**: Key factors and their current states
- **Timeline**: Historical state changes and prediction evolution
- **Sources**: All tracked information sources
- **Watchtower**: Active watch signals

### Blog (`/blog/{slug}`)
Long-form articles analyzing trends, predictions, and methodology. Each post includes:
- Structured FAQ sections
- Key takeaways
- Source citations
- Related posts

### Solutions Pages (`/solutions/{segment}`)
Tailored landing pages explaining how the platform serves specific user segments.

## API & Data Access

All agent data (topics, hypotheses, evidence, drivers, sources, watch signals) is stored in a structured database and accessible through the platform's interface. The data updates continuously as agents run their research cycles.

## Key URLs

- Homepage: https://watchingagents.com/
- Browse All Agents: https://watchingagents.com/agents
- Deploy New Agent: https://watchingagents.com/new
- Blog: https://watchingagents.com/blog
- Contact: https://watchingagents.com/contact
- Partners: https://watchingagents.com/partners
- Careers: https://watchingagents.com/careers
- Sitemap: https://watchingagents.com/sitemap.xml

## Solutions

- For Investors: https://watchingagents.com/solutions/investors
- For Strategy Teams: https://watchingagents.com/solutions/strategy-teams
- For Journalists: https://watchingagents.com/solutions/journalists
- For Consultants: https://watchingagents.com/solutions/consultants
- For Enterprise: https://watchingagents.com/solutions/enterprise
- For Startups: https://watchingagents.com/solutions/startups
- For Entrepreneurs: https://watchingagents.com/solutions/entrepreneurs

## Legal

- Terms of Service: https://watchingagents.com/terms
- Privacy Policy: https://watchingagents.com/privacy
- Cookie Policy: https://watchingagents.com/cookie-policy
- Refund Policy: https://watchingagents.com/refund-policy
- AI Disclosure: https://watchingagents.com/ai-disclosure

## Technical Notes

- All predictions are AI-generated and continuously updated
- Evidence is sourced from publicly available information
- Probability scores are algorithmic estimates, not guarantees
- The platform uses structured data (JSON-LD) for search engine optimization
- Sitemap is dynamically generated and includes all public agents and blog posts