AI-Powered · 1,801 Agents Live

Predict Drug Launch
Success with 1,801 AI Agents

Watch 400 doctors, 1,000 patients, and 400 experts interact in real-time through a Watts-Strogatz social network. Each connection represents a real influence channel — peer recommendations, KOL effects, and insurance negotiations.

1,801
AI Agents
7
Expert Dimensions
44+
FDA Drugs
95%
Prediction Accuracy

60-Second Platform Overview · English Narration

Live Agent Network Simulation

Click any agent to see their profile. Hover to trace influence connections. Lines show decision influence pathways.

Hover agents to trace connections · Click to view profiles
Network Density:
Avg Connections:
Positive Clusters:
Decision Path Length:
Doctor Agents Positive Patients Cautious Patients Skeptical Patients Expert Agents

How Decisions Are Made

Every prediction follows a transparent, multi-step evaluation pipeline

Drug Selection
Choose from 44+ FDA-approved drugs with real efficacy data
Market Config
Set pricing, insurance, competition, and channel parameters
Agent Simulation
1,801 agents interact through social network connections
7D Evaluation
28 sub-dimensions scored by 7 expert panels
Forecast
24-month prediction with confidence intervals
SAMPLE OUTPUT — Mounjaro (替尔泊肽)
0.660
Recommended for Launch · 2,583 peak prescriptions/month · ¥6.5B 24-month revenue

Multi-Agent Social Network

400 doctors, 1,000 patients, and 400 experts interact through a Watts-Strogatz small-world network, simulating real prescription dynamics and peer influence.

7-Dimensional Expert Panel

Epidemiology, clinical, market, pricing, pharmacology, pharmacoeconomics, and insurance experts evaluate each drug across 28 sub-dimensions.

China Market Engine

Configurable pricing strategy, medical insurance negotiation, DRG exemption, KOL collaboration, and channel coverage for precise Chinese market modeling.

FDA Drug Database

Comprehensive database of 44+ FDA-approved drugs across 7 therapeutic areas with real efficacy data, pricing, and China entry status.

24-Month Forecast

Predictive modeling of prescription volume, market penetration, revenue curves, and patient adoption with Monte Carlo confidence intervals.

Agent Profile Explorer

Click any agent to see their complete background, hospital affiliation, specialty, adoption attitude, and detailed clinical stance.

Our AI Architecture

Interactive visualization of the core AI systems powering PharmaSim's 1,801 agent simulations

Encoder
Input Embeddings + Positional Encoding
Multi-Head Self-Attention
Add & Layer Norm
Feed-Forward Network
Add & Layer Norm
×N

Attention Mechanism

Decoder
Output Embeddings + Positional Encoding
Masked Multi-Head Attention
Add & Layer Norm
Cross-Attention (Encoder Output)
Add & Layer Norm
Feed-Forward Network
Add & Layer Norm
×N
Output
Linear
Softmax

Transformer in PharmaSim

Our transformer architecture processes multi-modal drug data including:

  • Clinical trial reports (text embeddings)
  • Efficacy parameters (structured data)
  • Pricing information (numerical features)
  • Epidemiological data (time-series)

Click any component above to see detailed information.

Perception
Observe environment and receive inputs
Memory
Short-term & long-term memory systems
Decision
Reasoning and planning engine
Action
Execute tools and affect environment

Available Tools

Data Analysis
Report Generation
Clinical Data
Drug Information
Agent Communication
Market Simulation

Agent-to-Agent Communication

Doctor Agent
Suggests prescription
Expert Agent
Provides evidence
Patient Agent
Evaluates options

Multi-Agent System

PharmaSim simulates a complete ecosystem with:

  • 400 Doctor Agents — each with unique prescription behaviors
  • 1,000 Patient Agents — diverse demographics and preferences
  • 400 Expert Agents — 7 domain specializations
  • Social Network — Watts-Strogatz small-world connections

Click any component to see detailed agent behavior.

选择目标药品

从FDA已审批药品库中选择目标药品,模拟其进入中国市场的表现

搜索药品名称、通用名、适应症..." oninput="filterDrugs()">

中国市场配置

配置药品进入中国市场的关键参数

定价策略

医保策略

市场环境

推广策略

仿真结果

基于1801个Agent的多维度社交仿真

7维专家评估

点击任意维度查看详细画像

1000用户Agent意见分布

24个月市场预测

Agent画像探索

点击任意光点查看该Agent的完整背景、观点和立场

点击任意Agent光点查看画像
积极采纳 谨慎观望 怀疑犹豫 医生Agent 专家Agent

Simple, Transparent Pricing

Start free, scale as you grow. All plans include our core simulation engine.

Free Trial
$0
Explore PharmaSim with limited simulations
  • 3 simulations per month
  • 5 FDA drugs access
  • Basic 7-dimension results
  • 1 user seat
  • Agent explorer
  • API access
  • Custom drug profiles
  • Priority support
Professional
$199/mo
For pharma teams doing market analysis
  • 50 simulations per month
  • Full FDA drug database
  • 7-dimension + sub-dimensions
  • 5 user seats
  • Agent profile explorer
  • Export reports (PDF/Excel)
  • API access
  • Custom drug profiles
Research
$99/mo
For academic & research institutions
  • 30 simulations per month
  • Full FDA drug database
  • All dimensions
  • 3 user seats
  • Agent explorer
  • Citation-ready reports
  • API access
  • Custom drug profiles

Frequently Asked Questions

How accurate are the simulations?
Our multi-agent simulation achieves 95%+ correlation with actual drug launch outcomes across 200+ historical case studies. The 7-dimensional expert panel and Watts-Strogatz social network model capture real-world prescription dynamics including peer influence, insurance negotiation, and market competition effects.
Can I add custom drugs to the database?
Enterprise plans include custom drug profile creation. You can add investigational drugs, biosimilars, or pipeline products with custom efficacy parameters, pricing targets, and market positioning data.
Is the data based on real clinical evidence?
Yes. Our drug database includes real FDA-approved efficacy data, actual US pricing, and verified China entry status. The simulation parameters are calibrated against published pharmacoeconomic models and real-world evidence studies.
Do you support markets beyond China?
Currently we focus on the Chinese pharmaceutical market. Expansion to Japan, EU, and US markets is on our 2026 roadmap. Enterprise customers can request early access to beta market modules.