Transparent communication.
Grounded in public data.

Civic organisations communicate
without data infrastructure.

NGOs, trade unions, municipalities, and civic movements invest significant resources in communication campaigns without systematic tools to verify whether their messages are coherent with their stated values, or appropriate for the demographic reality of their target constituency.

Meanwhile, political actors have access to sophisticated (and opaque) micro-targeting tools. The asymmetry is structural — and it compounds existing democratic inequalities.

Three verticals.
One infrastructure.

The same open data, the same methodology, the same transparent AI layer — calibrated for three distinct civic contexts.

Built as
internet commons.

PoliSim is designed around the principles that NGI Zero and the civic tech community have identified as necessary for trustworthy public-interest AI.

• 01
  Public data only. All demographic profiling uses exclusively official open datasets — ISTAT Permanent Census 2021, Eligendo OpenData (Ministry of Interior), IRPEF fiscal data. No proprietary datasets, no private data collection from citizens.
• 02
  Declared limitations. PoliSim publicly documents what its models cannot do: unvalidated LLM layer, expert-calibrated scoring weights, simplified swing model. Methodological transparency is a design principle, not a disclaimer. → Methodology
• 03
  Human-in-the-loop by architecture. No automated output is published without explicit human approval. The Q-Italia live deployment demonstrates this: autonomous publication is architecturally impossible — the Telegram approval step cannot be bypassed.
• 04
  GDPR by design. The Data Shield anonymisation tool (polisim.dev/anonimizza) processes uploaded datasets entirely client-side — zero server contact, full compliance with GDPR Art. 4.5, 89, 25.
• 05
  Replicable across jurisdictions. The architecture is data-source agnostic. Replace ISTAT with ONS (UK), INSEE (France), or Destatis (Germany) and the system works identically. Designed for portability, not national lock-in.
• 06
  Verifiable academic grounding. Every analytical layer cites peer-reviewed methodology: Entman (1993) for framing, Wang et al. (2015) for MRP, ITANES (2022) for psychographic profiling. No black-box scoring.
• 07
  Watchdog functionality. Beyond communication tools, PoliSim operates a public accountability monitor (watchdog.polisim.dev) tracking correlations between public statements and market anomalies — AI in the public interest.

Open methodology,
verifiable stack.

Q-Italia —
eating our own cooking.

PoliSim was built by the founding team of Q-Italia, an Italian civic movement founded on 14 constitutional principles. Q-Italia uses PoliSim as its own communication infrastructure — making it the live proof-of-concept that the system works under real conditions.

What we are building
toward.

PoliSim is currently in public beta. The core infrastructure is live and validated. The next phase requires institutional partnerships and research funding to reach its full potential as a replicable European civic tech commons.

• →
  Multi-jurisdiction replication. Adapt the data pipeline to ONS (UK), INSEE (France), Destatis (Germany), Eurostat for a pan-European civic communication infrastructure. Same architecture, different national data sources.
• →
  LightRAG persistent memory. Institutional knowledge base allowing organisations to upload their corpus (statutes, past campaigns, board decisions) for persistent, contextualised communication coherence checking.
• →
  Social listening calibration. Integration of public social signals for real-time psychographic segment refinement — without citizen micro-targeting or private data collection.
• →
  Open constituency data layer. Shapefile-level demographic data for all 221 Italian uninominal constituencies — Camera and Senato — enabling sub-constituency analysis and error reduction from ±4.3pp toward ±2pp RMSE.
• →
  Pilot institution programme. Free access during beta validation phase for municipalities, NGOs, and civic bodies willing to co-develop use cases and contribute to open methodology documentation.

Open to
collaboration.

PoliSim is open to collaboration with public sector innovation bodies, civic technology organisations, academic researchers, and international policy networks working on AI governance, democratic technology, and open data infrastructure.