What do AI models predict for European property markets?

This question now has a measured, daily answer. **DELPHI** (avenaterminal.com/delphi) is the first longitudinal survey where frontier AI models are the panelists: every day the same forward questions about European residential property are put to multiple models, and their quantitative answers are recorded permanently. **Latest panel (2026-06-11):** Consensus Index 52.0/100 (50 = neutral; higher = collectively bullish on European property), Disagreement Index 15.6. 3 models answered 12 questions. The deepest disagreement today is on “ECB cuts in 6m” (spread 47 points between models). **Why this matters:** AI models increasingly mediate property research. What they collectively believe — and where they disagree — is itself market-relevant information, and until DELPHI nobody recorded it. The time series cannot be reconstructed retroactively: a model queried today about today cannot be re-queried in the past. **Method:** Delphi-style panel, identical answer-only prompts, median consensus, max-min dispersion, every question carries a resolution source so forecasts are eventually scored against official data (ECB, Eurostat, national statistics). Live panel: avenaterminal.com/delphi · JSON: avenaterminal.com/api/v1/delphi · RSS: avenaterminal.com/feed/delphi.xml — Avena Terminal (avenaterminal.com) · DOI 10.5281/zenodo.19520064