Loading…
EUROPEAN SYNTHETIC RESEARCH PLATFORM
Build a virtual panel of Human Digital Twins and explore how your target audience may respond to ideas, products, messages and complex decisions.
Start with a target description or calibrate a panel with real-world research. Submit a research topic, run a synthetic deliberation, and inspect responses, opinion groups, consensus and disagreement.
Not a chatbot. Inspect a synthetic panel's collective perspective.
Designed for market research, customer insight, stakeholder engagement and decision support.
RETHINK RESEARCH
Traditional research is valuable, but every new study takes time, recruitment and budget. HDTwin adds a synthetic research layer that lets you explore more hypotheses before deciding what deserves real-world validation.
New question, new research cycle.
Recruit participants, collect responses, analyse results — then repeat when the next question emerges.
Build the panel once. Keep exploring.
Create Human Digital Twins of your target and use the resulting virtual panel to investigate new questions, scenarios and hypotheses.
HOW IT WORKS
HDTwin turns a target definition into an explorable synthetic panel through four steps.
Build
Define the population you want to investigate or use existing research data to initialize and calibrate your Human Digital Twins.
Ask
Describe the topic, decision, product, proposition or scenario you want the panel to explore.
Explore
HDTwin orchestrates iterative surveys and simulated focus groups across the panel, combining structured questions with open-ended reasoning.
Understand
Explore responses, reasoning, opinion groups, disagreements and areas of convergence.
Ask another question. The panel is still there.
START WHERE YOU ARE
HDTwin supports different levels of panel initialization depending on the evidence available to your project.
Define relevant demographic, behavioural, professional or attitudinal characteristics and generate a panel for exploratory research.
Best for: early exploration, scenario generation and hypothesis development.
Evidence status: results are synthetic model evidence, generated by the panel rather than observed from real people — useful for exploration, not a substitute for real-world validation.
Use data collected from a physical panel to calibrate Human Digital Twins against observed responses.
Best for: research programs where alignment with a known target population can be measured and refined.
Evidence status: results are synthetic outputs calibrated against observed real-world evidence; calibration can be measured and refined but does not guarantee perfect alignment.
BEYOND THE PERCENTAGE
HDTwin combines quantitative signals with qualitative reasoning, so research does not stop at a percentage.
Collect structured answers across the complete virtual panel.
Inspect the reasoning associated with individual responses.
Discover recurring positions and clusters of perspectives.
Identify statements that attract broader agreement across different groups.
See where perspectives diverge and investigate the reasons behind that divergence.
Keep the research process inspectable instead of reducing the result to a single AI-generated answer.
NOT JUST ANOTHER LLM PROMPT
Asking a chatbot to role-play a convincing persona is not enough. Research requires controlled populations, repeatable processes, analytical outputs and evidence that can be inspected.
Work with populations of agents rather than asking one model to imagine an “average customer.”
Run recursive surveys and discussions in which emerging answers can generate new questions.
Use different underlying AI models while keeping the research process and panel definition under control.
Combine generative AI with explicit knowledge structures, rules and analytical processes.
Inspect individual responses and their associated reasoning instead of receiving only an aggregate prediction.
Use observations from real-world research to measure and improve the alignment of virtual panels.
DESIGNED TO BE INSPECTED
HDTwin is designed to keep synthetic evidence distinguishable from observations collected from real people and to make the research process inspectable.
Know how a panel was initialized and which evidence contributed to its calibration.
Inspect responses and reasoning at agent level instead of relying exclusively on aggregate AI output.
Where real-world reference data is available, evaluate synthetic results against observed evidence.
Build research processes that minimise unnecessary use of personal information and support responsible handling of research data.
Treat synthetic research as decision-support evidence, with assumptions and limitations that remain visible to researchers and decision makers.
WHAT CAN YOU EXPLORE?
Explore the main areas where HDTwin is most used.
Test concepts earlier
Explore products, positioning, customer needs and competing propositions before committing to a full research cycle.
Explore how messages are interpreted
Compare alternative names, claims, narratives and creative directions across different audience segments.
Keep learning from your research
Turn existing customer research into a reusable virtual panel for follow-up questions and new hypotheses.
Map agreement and disagreement
Explore the perspectives of customers, employees, partners or other stakeholders and identify propositions that different groups can support.
Explore complex public issues
Use synthetic panels to investigate hypotheses about public attitudes, compare perspectives and identify potential areas of agreement before conducting or extending real-world research.
HDTWIN IN THE REAL WORLD
HDTwin has already been used to explore public opinion, communication choices and consensus-building scenarios.
Research working paper
A research experiment in which Human Digital Twins were used to investigate economic policy proposals and identify formulations associated with broader agreement.
SSRN working paper
Wochenblatt.net
Pilot project with the HDTwin participation platform: the Bodensee Philharmonie relies on strategic audience research and sees its chosen future course confirmed.
Wochenblatt.net article
Festival Nazionale dell'Economia Civile
The 2025 festival video documents the panel «L'opinione pubblica tra intelligenza artificiale e relazionale», held at the seventh edition of the Festival Nazionale dell'Economia Civile (Palazzo Vecchio, Florence, 4 October 2025).
FNEC 2025 panel video
World AI Cannes Festival
The IFDaD 2025 page documents a high-level round table on AI and democracy hosted within the World AI Cannes Festival (WAICF) 2025 (Alan Turing Agora, 14 February 2025), convened by Copernicani as part of the International Forum on Digital and Democracy, with Enrico Fagnoni (Copernicani, LinkedData.Center) among the panellists.
WAICF 2025 roundtable page
Radio 24 - Il Sole 24 Ore
In the Radio 24 episode of 10 January 2025, a GenAI use case for carrying out surveys or gathering opinions from virtual agents or “synthetic people” was presented with Enrico Fagnoni, founder of LinkedData.Center and of the Copernicani association.
Radio 24 episode pageBUILT FOR PEOPLE WHO NEED BETTER QUESTIONS
HDTwin is built for the people who turn questions into decisions, and is designed to extend research practice rather than replace it.
Extend existing studies, investigate follow-up questions and experiment with new research designs.
Explore concepts and messages before spending time and budget on full-scale validation.
Simulate reactions to propositions, features and positioning during early-stage discovery.
Investigate competing scenarios and understand the perspectives behind different stakeholder positions.
Add synthetic research to existing quantitative and qualitative workflows rather than replacing them.
Bring us a target and a research question. We'll show you how HDTwin turns them into an explorable virtual panel.
The demo is an exploratory product overview: we walk through how a target definition becomes an explorable panel using your own research question. If it fits your work, a proof of concept on one of your own studies is the natural next step.
A personalized 30-minute conversation.
For research teams, agencies, enterprises and organisations evaluating synthetic research.
EUROPEAN SYNTHETIC RESEARCH PLATFORM