2026-03-30 20h — Creative, Cross-border & Unconventional

2026-03-30 20h — Creative, Cross-border & Unconventional

The Problem

European real estate owners face a compliance tsunami. The EU's new Integrated Framework for Climate Resilience (expected Q4 2026) introduces "climate resilience by design" — meaning every new infrastructure investment must demonstrate inherent resistance to projected climate impacts. Meanwhile, EU Taxonomy revisions (Q2 2026) tighten technical screening criteria for climate adaptation.

Property owners — especially mid-market firms managing 10-500 properties — have no simple, affordable tool to: (1) assess physical climate risk per asset (floods, wildfires, heatwaves, coastal erosion), (2) quantify financial impact on asset value and insurance premiums, and (3) generate compliance reports for EU Taxonomy, CSRD, and SFDR.

Current options are either expensive consultant engagements (€20K-€100K per portfolio assessment) or complex enterprise platforms designed for the largest institutional investors.

The Problem

European SMBs are deploying AI agents at an unprecedented rate — from customer service bots to coding assistants to autonomous data processors. By end of 2026, Gartner predicts 40% of enterprise apps will feature task-specific AI agents. But these agents operate with zero identity governance:

• No verifiable identity: Who deployed this agent? What permissions does it have? What did it do?
• Excessive privileges: AI agents routinely get admin-level API keys that far exceed what they need
• Zero audit trail: When an AI agent accesses customer data, sends an email, or makes a purchase — there's no identity-aware log
• EU AI Act compliance gap: From August 2, 2026, AI systems need transparency and governance; but IAM tools only manage human identities

Existing NHI solutions (Oasis Security, Entro, CyberArk) target enterprises with 10,000+ identities and six-figure contracts. European SMBs deploying 5-50 AI agents have no affordable option.

The Problem

Portuguese is the world's 6th most spoken language with 260M+ native speakers across 9 countries. Yet Portuguese + Spanish combined make up only 2-3% of AI training material. The result: AI models consistently underperform for Portuguese speakers — poor text generation, inaccurate speech recognition, culturally inappropriate outputs, and weak understanding of PT-specific contexts.

This data gap creates three pain points:

1. AI companies building multilingual products can't find high-quality Portuguese training data (especially European Portuguese, African variants, and domain-specific content)
2. Portuguese companies adopting AI tools get inferior results compared to English-language counterparts
3. CPLP governments and institutions investing in AI (Portugal's AI Portugal 2030, Brazil's AI strategy, Angola's digitalization push) lack local data infrastructure

Existing data providers (Scale AI, Appen) treat Portuguese as one of 100+ languages — no specialization, no cultural depth, no CPLP-wide coverage.