LEARNING
AI Agent Design: Build Teams, not Bots
Learn how to design AI agents as collaborative digital colleagues using unified platforms to enhance team cohesion, productivity, and trust.
Imagine a global council of experts debating UBI in Finnish, Hindi, Japanese, and Arabic. Our AI experiment brings authentic cultural reasoning.
When One Perspective Isn't Enough
In a world grappling with complex, interconnected challenges—from AI governance to climate policy—single-perspective analysis falls short. The most pressing questions of our time demand diverse viewpoints, cultural nuance, and genuine intellectual debate. But assembling experts from around the globe for real-time deliberation? That's expensive, time-consuming, and often logistically impossible.
What if you could convene a global council of experts in minutes, each reasoning in their native language, each bringing authentic cultural context to the table?
Introducing Vertesia's Expert Roundtable
Vertesia's Expert Roundtable feature transforms how organizations approach complex decision-making. It orchestrates multi-round debates between specialized AI agents—each with distinct perspectives, personalities, and even different underlying models—to produce nuanced, bias-resistant analysis.
Unlike traditional AI tools that provide a single viewpoint, Roundtable creates a deliberative democracy simulation where diverse perspectives clash, converge, and synthesize into richer insights than any single agent could produce alone.
Key capabilities:
- Multi-model diversity: Deploy different AI models (Claude, Gemini, GPT) to ensure genuinely different reasoning patterns
- Cultural authenticity: Agents can debate in their native languages, surfacing culturally-embedded concepts
- Dynamic evolution: Positions evolve across multiple rounds as agents respond to counterarguments
- Intelligent synthesis: A moderator agent orchestrates the debate and produces unified recommendations
A Real-World Demonstration: The UBI Global Perspectives Council
To showcase the power of this approach, we ran an extraordinary experiment: Should a Global AI Tax fund Universal Basic Income?
We assembled a virtual council of four experts, each representing a different nation, reasoning in their native language, and powered by different AI models:
- 🇫🇮 Finland (Claude Sonnet, Finnish): Social safety net advocate
- 🇮🇳 India (Gemini Pro, Hindi): Emerging economy perspective
- 🇯🇵 Japan (Claude Opus, Japanese): Work ethic skeptic
- 🇦🇪 UAE (Gemini Pro Preview, Arabic): Resource-wealth model advocate
Round 1: Authentic Cultural Perspectives Emerge
The opening statements revealed genuinely different worldviews—not translated English concepts, but culturally-rooted reasoning:
Finland cited their 2017-2018 UBI experiment showing improved mental health and increased work participation. They argued UBI represents civilization's "floor," not luxury—a perspective rooted in Nordic social democracy.
India championed the JAM trinity (Jan Dhan, Aadhaar, Mobile) as proof that digital infrastructure enables direct transfers to 1.4 billion people, bypassing corruption. They framed UBI as "power transfer," not charity—reflecting India's experience with financial inclusion at scale.
Japan opposed universal UBI as threatening ikigai (life purpose). They cited 1.46 million hikikomori (socially isolated individuals) as evidence of a purposelessness crisis. Their "Ikigai Economy" proposal invested 70% in education and participation, only 30% in safety nets—embodying Japanese work culture values.
UAE advocated for a sovereign wealth fund model: invest AI tax revenues and distribute only dividends, not capital. Citizens should be "shareholders, not recipients"—a perspective born from Gulf states' experience managing resource wealth.
Round 2: Cross-Pollination and Evolution
The moderator relayed counterarguments between experts. Japan challenged Finland's "passive income" model. India questioned UAE's equity approach for the unbanked. Finland pressed Japan on immediate needs during automation crises.
Something remarkable happened: positions began to evolve. Japan acknowledged the need for safety nets if paired with purpose-building. UAE recognized the urgency of immediate transfers for vulnerable populations. Genuine synthesis was emerging.
Round 3: The Geneva Framework
By the final round, all four experts—despite starting from opposed positions—converged on a sophisticated three-pillar hybrid model they called the "Geneva Framework":
- Pillar 1 (40%): Immediate direct cash transfers to the poorest populations
- Pillar 2 (35%): Global AI Sovereign Wealth Fund with equity stakes in AI companies
- Pillar 3 (25%): Education, reskilling, and entrepreneurship (the "Ikigai Clause")
They proposed GAIRA (Global AI Revenue Authority) under UN oversight with blockchain transparency and a 3-year pilot in 20 countries.
The vote? Unanimous conditional approval (4-0).
What Makes This Extraordinary
This wasn't a parlor trick. The demonstration showcased capabilities that redefine how organizations can approach complex decisions:
1. Authentic Multilingual Reasoning: Each model argued in its native language, forcing culturally-embedded thinking. Concepts like ikigai, bayt al-mal (Islamic treasury), and JAM trinity emerged organically—ideas that might never surface in English-only analysis.
2. Model Diversity = Cognitive Diversity: Using four different models across two cloud providers (AWS Bedrock, Google VertexAI) created genuinely different reasoning patterns, not just rephrased perspectives.
3. Dynamic Position Evolution: Participants didn't just state positions—they evolved through structured debate, making concessions and finding common ground. Japan moved from principled opposition to conditional support.
4. Intellectual Rigor: Participants cited specific data (McKinsey estimates, IMF projections), used structured arguments, and proposed concrete governance mechanisms—producing policy-grade analysis.
5. Genuine Synthesis: The final framework wasn't a lowest-common-denominator compromise. It incorporated Finland's immediacy, UAE's investment wisdom, Japan's purpose-focus, and India's digital infrastructure into a coherent solution all could support.
Applications Across Industries
The implications extend far beyond policy debates:
Strategic Planning: Test business strategies against multiple market perspectives simultaneously—optimistic vs. skeptical, domestic vs. international, technical vs. commercial.
Risk Assessment: Evaluate decisions through security, legal, financial, and operational lenses in parallel, surfacing blind spots early.
Product Development: Gather diverse user perspectives—accessibility advocates, power users, enterprise buyers, developers—in a structured debate before committing resources.
Research & Analysis: Tackle contentious research questions by having different methodological perspectives (quantitative vs. qualitative, theoretical vs. applied) debate the evidence.
The Future of Decision Intelligence
Vertesia's Expert Roundtable represents a fundamental shift: from asking "What does AI think?" to orchestrating "What do diverse AI perspectives reveal when they debate?"
The result is decision intelligence that's more robust, less biased, and richer than any single perspective could provide. It's multicultural deliberative democracy at machine speed—and it's available now.
Want to see how it's done? Watch my video on x.com
See the full report
See the log of the agentic run
Ready to see what emerges when diverse perspectives collide? Discover how Vertesia's Expert Roundtable can transform your most complex decisions. Submit a demo request.