This article covers OriginalVoices, an AI startup, raising £1m in a pre-seed funding round to build a human data layer for generative AI. The funding aims to scale the startup's network of consented 1:1 digital twin profiles to support market research, personalised content and enterprise AI workflows by providing real-time human opinion.
OriginalVoices, an AI startup based in London, has raised £1 million in a pre-seed funding round to build a human data layer for generative AI. The platform creates and maintains 1:1 “digital twin” profiles owned by real people so AI agents and tools can query up-to-date human opinion and behaviour — a capability that aims to reduce the gap between model outputs and real-world audience insight.
As large language models produce content at scale, they still reflect patterns in their training data rather than current human attitudes about specific products, brands or topics. That mismatch can make AI outputs feel generic or misaligned with target audiences. OriginalVoices’ approach seeks to provide a direct channel to real people’s views, which could matter for product testing, market research, personalised content and compliance with provenance expectations in AI deployments.
OriginalVoices runs a consumer app called Twineo where tens of thousands of people create, train and manage their own digital twin profiles. Each twin is intended to represent a single, consenting person; users can update and validate their twins continuously and earn money when their twin contributes to research.
The network of twins is accessible on demand via the platform’s API and through integrations with larger AI environments, with the company saying agents can query profiles inside tools such as ChatGPT and Claude. OriginalVoices launched the network in December 2025 and has since added integrations and partnerships designed to turn static panel data into interactive, real-time respondents.
A notable partnership is YouGov Parallax, a joint product that pairs YouGov’s two-decade-old survey panel with OriginalVoices’ twin infrastructure to let customers query “living” twins drawn from established polling samples. The arrangement illustrates how legacy research firms and newer data-layer startups can combine audiences and technology to serve enterprise customers.
The pre-seed round was led by Iona Star and totals £1 million. The funding will be used to grow the team, expand the digital twin network, build further partnerships, sharpen the product and push into the US market, while refining the platform’s real-time human data for improved audience targeting and AI personalisation.
Iona Star is an early-stage investor with recent activity in AI training data and foundational AI projects; the firm has backed Yann LeCun’s AMI Labs and led a previous round in Worldmodeldata.
In the announcement, Kevin McGivern, Partner at Iona Star, said:
Across industries, the limit to AI effectiveness is no longer model capability; it's the quality of the data feeding it. OriginalVoices stands out by bringing a proprietary, real-time human insight layer to AI workflows that reflects the nuances and realities of human opinion. That's a structural advantage, and we want to be part of supporting OriginalVoices' development to improve the future of AI.
If you're researching potential backers in this space:
In the announcement, David Dobrin, Co-founder & CEO at OriginalVoices, said:
Choosing an investment partner isn't easy, but Iona Star brings a genuine understanding of the space we're building in and acts as an extension of our team, which we are extremely grateful for. As AI models get more powerful and can generate almost anything instantly, the real context gap is human insight, opinion and feedback: knowing what people think and feel, not just relying on a model guessing or making assumptions from its training data. That's what OriginalVoices gives - confidence that AI decisions and outputs are grounded in real human perspective and audience understanding - leading to improved decision-making and outputs. With Iona Star's backing, we are ready for our next phase: growing the team, sharpening our product offering, and expanding further into the US. Dobrin’s statement reaffirms the startup’s priorities: expand the twin base, deepen integrations with AI tools and convert partnerships into paying customers.
OriginalVoices sits at a junction of two trends: demand for human-in-the-loop data to improve AI outputs, and legacy research firms seeking ways to make panels interactive for modern AI workflows. The YouGov tie-up is an example of incumbents leaning on startups for technical infrastructure while contributing curated audiences.
For the UK and wider European AI ecosystem, the deal highlights continued investor interest in data infrastructure rather than model architectures alone. As regulation and customer expectations increasingly emphasise provenance and human oversight, startups that can reliably connect models to living, consented audiences may find a growing market among enterprise customers and research buyers.
| Investors | Investment Focus | Startup Investments | Round Size | Connect |
|---|---|---|---|---|
![]() Iona Star( ) Belfast | ||||
| All investors | All investor sectors | All funded startups | All funding rounds |
Click here for a full list of 7,589+ startup investors in the UK