Metaview

Company overview

Metaview, founded in 2018 by ex-Uber and ex-Palantir leaders Siadhal Magos and Shahriar Tajbakhsh and headquartered in London (with a US entity, Metaview Technologies Inc.), began as an AI interview notetaker that records, transcribes and summarizes recruiting interviews and generates structured notes and scorecards. It has since expanded into a broader recruiting platform: sourcing, candidate search, job posts, reporting, and an Application Review agent that turns a job description into an Ideal Candidate Profile and screens every applicant into 'Strong / Good / Partial match' buckets with written rationale. The company has over 2,500 customers (including Sony, Brex, Deel and Deliveroo) and raised a $35M Series B led by GV in 2025.

Regulatory exposure

While the AI Notetaker is positioned as non-decisioning ('doesn't make decisions, evaluate candidates, or provide recommendations'), the Application Review product ranks and rates candidates against a profile, which makes it an automated employment decision tool implicating NYC Local Law 144 and likely EU AI Act Annex III high-risk scrutiny. Metaview engages Warden AI for monthly, ongoing bias audits of Application Review (latest 7 May 2026, 15,264-candidate sample) with a publicly downloadable NYC LL 144 report and intersectional sex/race analysis. On the EU AI Act, Metaview publicly asserts a non-high-risk position under the Article 6(3) exemption rather than publishing high-risk-deployer (FRIA) documentation. It is SOC 2 Type II certified, GDPR/CCPA compliant, hosts data in AWS UK, and states it does not train AI models on customer data.

Path to a higher score

Metaview already clears the hardest bar (continuous, publicly downloadable third-party bias auditing) and has strong security/data-governance disclosure. To raise its scores it should: publish Article 11-style technical documentation or a model card / explainability statement for Application Review and a responsible-AI policy; provide explicit NYC LL 144 deployer guidance and candidate-notice templates rather than leaving compliance to customers; document its EU AI Act Article 6(3) reasoning and offer FRIA / deployer-obligation guidance for customers who do treat it as high-risk; surface a public sub-processor list and a status/incident channel; and pursue ISO 42001 to formalize AI governance.