Poindexter Labs raises £2M to improve training data for advanced AI

UK-based Poindexter Labs, an AI data company focused on producing expert-level training data for advanced AI models, has raised £2 million in an oversubscribed seed round. The funding was led by Episode 1 and First Cheque, with participation from partners at Venture Friends and a number of angel investors. The round also included investments from several of Poindexter's own contributors, including mathematicians and scientists who work on the platform.

As AI models become increasingly capable of tackling complex reasoning tasks, demand is growing for high-quality training data created by domain experts. However, much of the industry's existing data infrastructure was built for simpler annotation tasks such as image labelling and text classification.

Advanced AI systems require something different: detailed expert reasoning, professional judgement, and domain-specific knowledge across fields such as mathematics, science, law, medicine, finance, and engineering.

Poindexter has translated its methodology into a proprietary platform, currently in beta, that supports both its data production services and organisations developing their own AI systems. Through the platform, enterprises and government departments can access collaborative workflows designed to help subject-matter experts generate and validate training data more effectively.

According to Jocelyn D'Arcy, Founder and CEO of Poindexter Labs, many of the workflows still used to generate AI training data prioritise scale over collaborative knowledge creation.

As a result, a huge chunk of training data is discarded not because it is wrong, but because adversarial review processes actively incentivise discarding tasks rather than improving them. This is a workflow problem, and we built Poindexter the way academics build knowledge: collaboratively, transparently, with peer review at every step.

Poindexter Labs plans to use the funding to scale its team, deepen relationships with frontier AI labs, and bring its platform to a broader range of enterprise and public-sector organisations developing AI systems.