Proxion Launches to Build the Financial Judgment Infrastructure for the Next Generation of AI

New York, USA, Oct. 21, 2026 (GLOBE NEWSWIRE) -- Proxion, a New York-based AI data and evaluation company, today announced its official launch, six months after its founding, with a network of more than 1,000 senior finance professionals and commercial engagements with AI labs and financial AI companies.
Proxion is working to make the expertise and scrutiny of institutional finance a measurable part of AI development. Its rapid growth reflects increasing demand for specialized financial training data and evaluations.
AI is moving beyond retrieving information and summarizing documents to analyzing transactions and executing complex financial workflows. This shift demands more than technical accuracy. AI systems must demonstrate that their conclusions can withstand institutional scrutiny.

“As AI takes on increasingly complex financial work, an accurate answer is only the starting point,” said Taron Arshakian, Founder and CEO of Proxion. “In finance, you need to know which assumptions to challenge, how to assess uncertainty, and when the numbers don't tell the full story. Those are the capabilities we need to develop, test, and measure.”
Proxion combines senior financial expertise with structured quality controls to produce specialized training data and evaluations. Experienced professionals develop reference analyses, reasoning traces, and detailed scoring frameworks, supported by independent review. This allows AI systems to be evaluated not only on their final outputs, but also on the evidence, assumptions, and reasoning behind them.
An AI-generated financial model can contain fully integrated financial statements and attractive investment returns while relying on unsustainable leverage, unsupported EBITDA adjustments, or financing terms inconsistent with market conditions. The calculations may be correct while the transaction itself remains commercially unviable.

“The more convincing an AI-generated analysis looks, the harder its underlying mistakes can be to detect,” Arshakian said. “As AI systems take on greater financial responsibilities, those mistakes can propagate across entire workflows. The challenge is making professional scrutiny systematic, measurable, and scalable.”
Proxion's long-term ambition extends beyond individual datasets and evaluations. As AI becomes increasingly embedded in institutional finance, the company aims to establish the benchmarks and continuous performance standards needed to measure reliability throughout development and deployment. As models and markets evolve, these standards will provide AI developers and financial institutions with a consistent foundation for understanding what AI systems can reliably do, where their limitations remain, and when human oversight is essential.
About Proxion
Proxion is a New York-based company building the financial judgment infrastructure for the AI era. Founded in 2026, Proxion transforms institutional finance expertise into specialized training data, evaluations, and benchmarks that advance the capabilities of frontier AI. Its mission is to establish the global standard for financial AI judgment, defining how intelligent systems are trained, evaluated, and ultimately trusted to operate in the world's most consequential financial environments.
Website: www.proxion.ai
X (formerly Twitter): X

Contact

Taron Arshakian
Proxion, Inc.

taron@proxion.ai

Attachment

• Proxion

Proxion

Proxion