iB3 SA builds AI platforms that simulate real populations, extract human intelligence, assemble optimal teams, and profile decision-makers. European-built, Swiss-hosted, privacy-first.
AI-powered market research with simulated populations. Create digital twins of real demographics using Eurostat and World Values Survey data, then run virtual focus groups, A/B tests, and policy simulations in days instead of months.
Beliefs-Desires-Intentions engine with credibility filtering, population variance modeling, and data-driven cognitive blocks. Validated against real survey data with 85% directional accuracy.
Cross-study knowledge store with pgvector for semantic retrieval across all past research
IPF-calibrated demographics across 71 profiles in 11 countries. 8 variables including occupation and household type. Schwartz SVS Wave 7 cultural priors per region.
7-phase autonomous research pipeline with explainability and provenance tracking
Qualitative deep-dive with inter-agent discussion and cluster formation
Full quantitative + qualitative panel with independent BDI analysis
Test 2-5 message/creative variants with Pixtral Vision AI
Single stimulus, standard BDI analysis in 2-3 minutes
Longitudinal tracking with same cohort over time
Compare 2-5 text variants on same cohort with ranking
Borda count ranking with Kendall's W agreement coefficient
Multi-lever factorial analysis for combination optimization
Longitudinal multi-study design with portfolio analysis and trend tracking
BDI cognitive pipeline, demographic calibration, adaptive wizard
Segmentation drivers, cognitive block analysis, KPI scoring
Pixtral Vision AI for automatic image analysis in A3 studies
RAG-powered cross-study knowledge accumulation with pgvector
Consultant workflow, client visibility toggle, investigation notes
Prioritized three-tier implementation plan with Messaging Playbook per segment
Focus Study CHF 4,400: 5,000 AE, max 200 agents, 3 levers, all 9 study types. Upfront CHF 3,000 / Balance CHF 1,400 (30 days).
Strategy Design CHF 12,000: 25,000 AE, max 500 agents, 6 levers, all 9 types, Decision Layer + P1-P6. Upfront CHF 6,000 / Balance CHF 6,000 (30 days).
Research Program CHF 19,000: 100,000 AE, max 1,000 agents, 8 levers, all 9 types, Decision Layer + Custom Demographics + P1-P6. Upfront CHF 9,000 / Balance CHF 10,000 (30 days).
1 AE = 1 virtual human × 1 lever. Free Preview available — design any study before purchase.
Focus CHF 2K/mo: 500 AE/month, max 200 agents, 3 levers
Strategy CHF 4K/mo: 2,500 AE/month, max 500 agents, 6 levers
Research CHF 7.5K/mo: 10,000 AE/month, max 1,000 agents, 8 levers
Digital twin simulation for youth athlete development. Practice real conversations with an AI twin of each athlete, extract BDI profiles from psychologist interviews and coach diaries, and track growth, team dynamics, and performance trajectories through a 1,400+ node knowledge graph.
Rehearse real conversations with an AI simulation of each athlete, grounded in their actual psychological profile — practice difficult conversations before you have them.
Extract Beliefs, Desires, and Intentions from psychologist interviews and coach diary entries
1,400+ nodes, 4,500+ edges with D3.js visualization and semantic search explorer
5-tab dashboard: physical tests (vIFT, NordBoard, ForceDecks, SmartSpeed), daily wellness, coach evaluations, psychological markers, correlation analysis
11 analytical views: Graph, Matrix, Alerts, Social Health, Gatekeepers, Frustrated Triangles, Scan, Cross-Diary, Q&A, Reports, Timeline
AI-powered political intelligence with virtual voter panels calibrated to real electorates. Five simulation types — Candidate Ranking, Ideal Candidate, Message Testing, Policy Testing, and Panel Q&A — with Schwartz+OCEAN personality modeling, cognitive block detection, and deep narrative analysis. 53 regions across 11 countries.
53 regions, 8 demographic variables, IPF calibration from Eurostat, Schwartz SVS Wave 7 priors. Same demographic rigor as Synthesight consumer panels.
Perplexity AI + Apify pipeline: professional profile, party alignment, media presence. Three trigger modes. Language-aware queries in Italian, German, and French.
Four politically-specific barriers: Trust Deficit, Fairness Block, Identity Threat, Status Quo Bias. Each with affected percentage and strategic recommendation.
Batch evaluation with Schwartz+OCEAN personality integration. All agents evaluated per candidate in a single LLM call (~30 seconds). Deep narrative analysis for all 5 study types.
Intelligent team composition for professional associations. Extract competencies from CVs, build BDI-based member profiles, and let AI assemble optimal teams with full GDPR compliance, TOTP MFA, and comprehensive PDF reporting.
Skill Match, Domain Fit, Experience Relevance, Availability, Cost Fit, Complementarity. Three optimized team proposals in 60 seconds.
Per-member justification with portfolio evidence, gap analysis, risk assessment, and market context. Professional PDF reports for board presentations.
Enterprise-grade security with time-based OTP and full GDPR data export/deletion compliance
Comprehensive reporting: team composition, member profiles, competency matrices, and assembly rationale
AI-driven jury selection intelligence for trial teams. Profile every prospective juror on Schwartz values, OCEAN traits, and case-specific BDI reasoning, generate targeted voir dire questions, rehearse how the panel will deliberate, and monitor every strike for Batson and GR37 fairness compliance.
Tiered psychological profiling (J1–J3) combining ten Schwartz values, five OCEAN traits, and case-specific BDI reasoning — enriched with intelligence sweeps and uploaded documents.
Verdict lean, leadership, holdout, malleability, deception, and Batson vulnerability — each scored and paired with a strike, keep, monitor, or cause-challenge recommendation.
Multi-round agent-based simulation predicts the verdict, foreperson, holdouts, and critical arguments — with what-if testing for any proposed strike.
Real-time monitoring of strike patterns against seated-juror demographics, with race-neutral justification prompts and a Batson-safety indicator for every juror.