Full Platform Capabilities

Profile Generation

AnthroSim profiles are built on a proprietary quantitative grounding model — not random persona generation. Where other tools pick traits arbitrarily, AnthroSim derives each profile from harmonized human survey and population data, then layers in contextual signals from the person’s environment, profession, and life circumstances. The result: personas that hold together internally, across economics, politics, psychology, and behavior, in ways that reflect how people actually are.

Example input: 10 profiles · Guidance: “urban professionals with views on housing policy”

Each generated profile reflects:

• Population-grounded demographics — sampled from real distributions, not uniform random selection
• Multi-dimensional political modeling — independent stances across a range of policy dimensions that don’t collapse into a simple left/right axis
• Economically consistent life circumstances — occupation, financial situation, and material security that hang together and shape how each persona sees the world
• Behavioral depth — psychological parameters that govern how personas actually respond to arguments, evidence, and pressure in conversation
• Narrative grounding — rich personal history that anchors each persona’s behavior beyond surface-level traits

Example profile — illustrative only. Production profiles are substantially richer internally than the compact cards shown on the site.

What AnthroSim delivers: The capability to generate, organize, and deploy persona cohorts in directed conversations. Generated profiles drive simulation behavior internally — they are not exported as standalone data artifacts by default. The same cohort can be run against multiple topics or experimental conditions with full reproducibility.

Profile data export — including underlying model parameters and individual profile internals — is available under commercial license. Contact us to discuss access.

Conversation Simulation

Run structured group discussions with 10–500 AI participants. Chats can run simply for fast iteration, or in orchestrated mode for richer, more realistic conversational dynamics.

Example input: Topic: “inclusionary zoning requirements” · Participants: 30 · Duration: 90 min

Example conversation excerpt:

Real-time data integration: Participants can fetch and cite current statistics via the Kagi API — recent policy changes, economic data, regional statistics — scoped automatically to the discussion topic.

Project materials: Attach participant-facing materials — briefs, case facts, product specs, discussion guides, or source packets — and optionally allow web research so simulations can draw from the same context your real audience would read or reference.

Voting & Analytics

After discussion, AnthroSim collects structured votes from all participants. Voting decisions are driven by each participant’s complete personality profile and the actual content of the conversation — not predetermined outcomes.

Example output:

{
  "vote_summary": {
    "topic": "Inclusionary zoning at 15% threshold",
    "total_participants": 47,
    "support": { "count": 27, "pct": 57.4 },
    "oppose":  { "count": 13, "pct": 27.7 },
    "neutral": { "count": 7,  "pct": 14.9 }
  },
  "representative_statements": {
    "support":  "Marcus Webb: 'With real enforcement teeth, I'm in.'",
    "oppose":   "Laura Kim: 'This kills small-scale development. Not the right tool.'",
    "neutral":  "Dr. Nair: 'Promising, but needs a 5-year review clause.'"
  }
}

Results are announced by the moderator as part of the natural conversation flow, then exported to JSON for analysis. Aggregate vote totals and directional breakdowns by demographic group are the most reliable outputs — treat fine-grained subgroup splits as signals for further investigation, not as statistically definitive measurements.