Unlocking Human Insight for AI: Our Investment in OriginalVoices

Artificial intelligence is advancing at unprecedented speed. Models are becoming faster, cheaper, and more deeply embedded into products, workflows, and everyday life.

Until now, almost all the data powering AI has been used to train foundation models. But we believe the next critical layer is high-quality consumer data and insight at inference - the kind that can tailor, shape, and improve outputs in real time.

Drawing on our deep knowledge of leveraging data across markets, Iona Star has invested in OriginalVoices to help it deliver its thesis that the quality and effectiveness of AI workflows and outputs will increasingly depend on access to unique, proprietary real-time data, to enable deep personalisation and precise audience targeting.

An AI-native Approach to Human Insight

Traditional research methods weren’t designed for AI-native workflows. They are static, expensive, and slow - taking days or weeks to generate insight for systems operating in milliseconds.

AI tools and applications need insight that is fully customisable, available instantly, but still grounded in real people. A new approach is required - one that fuses the power of AI with real human input.

A Human Intelligence Layer for AI

OriginalVoices is building a new layer in the AI stack: a Human Intelligence API that gives AI systems instant access to human insight. At the core is a network of 1:1 Digital Twins: AI representations of individuals that are owned, trained, and validated directly by the people they represent.

Through the OriginalVoices API, MCP, and Self-Serve platform, customers can ask any question to any audience - and receive instant, human-grounded insights that feed directly into prompts and workflows to inform, shape and improve outputs.

“Foundation models are incredibly powerful and can generate almost anything instantly - but they don’t know what real people think or care about. That means applications and agents building on top of them lack the human context required to differentiate outputs and make them truly relevant. OriginalVoices bridges that gap by providing real-time human insight at the point of inference to guide, shape, and improve outputs,” said David Dobrin, Co-Founder and CEO of OriginalVoices. “We built OriginalVoices because we believe that in the age of AI, what real people think and feel matters more than ever. Our vertically integrated approach - building both a consumer Digital Twin network and an enterprise platform - ensures that real people are rewarded for their views, while making their collective opinions and insights accessible instantly and on demand.”

“Across industries, the limit to AI effectiveness is no longer model capability - it’s the quality of the data feeding it,” said Kevin McGivern, Partner, Iona Star. “OriginalVoices stands out by bringing a proprietary, real-time human insight layer to AI workflows. That’s a structural advantage. It allows its customers to generate messaging and content that’s not only targeted but provides an unprecedented view into what real audiences think and feel in real time. We backed the team because we see this as a foundational layer for next-generation AI applications.”

Why This Matters

OriginalVoices sits between foundation models and real people - providing a dynamic human insight layer at the exact moment AI tools generate outputs.

We see an opportunity in ensuring that intelligent systems understand and reflect the nuance and reality of human opinion - not just the averages of the historical internet.

The team is building the data infrastructure that makes AI not just generative - but genuinely aligned with real people and real audiences.

With thousands of Digital Twins active and OriginalVoices now released, we’re proud to support them in bringing that future to life.