In Profile: Thomas Galbraith at Barker

Fintech firm Barker is leveraging proprietary artificial intelligence and insurance partnerships to fundamentally alter how financial institutions approach asset-backed lending, especially for hard-to-price collateral like fine art and industrial equipment. The company provides real-time valuations for assets and backs that price with a warranty, shifting the downside risk from the lender to the insurer.

The inspiration for Barker, according to Co-Founder Thomas Galbraith, was the persistent problem of price discovery for hard assets. Historically, traditional valuation methods were manual and analog, often yielding disparate results that failed to inspire confidence in lenders. Furthermore, these traditional firms rarely guaranteed the accuracy of their assessments.

“Hard assets have always suffered from the same problem: price discovery happens at the point of resale,” Galbraith explained. “At Barker, we took a different approach. We use our own domain-specific LLM, or SLM, to accurately price assets. We then back up that price with a warranty such that if our price turns out to be wrong, we pay the difference. This has fundamentally changed the way valuations work in asset-backed lending.”.

Barker’s technology addresses the core issue of risk for lenders by warrantying the accuracy of its valuation outputs. If an asset is sold for less than the price determined by the AI, Barker and its A-Rated insurance partners, such as Munich Re, cover the shortfall.

“Effectively, by warrantying the accuracy of our outputs, we are assuming the risk of our being wrong, instead of the lender assuming all the risk,” Galbraith commented. He noted that the warranty is straightforward: if Barker values an asset and it sells for less, the company is responsible for the error and pays the difference between the sale price and the initial valuation.

The AI Valuation Mechanism

To value such diverse asset classes, Barker’s system treats asset pricing as a series of interlinked cognitive tasks. The company’s master Large Language Model (LLM), which is generalised for heterogeneous physical asset valuation, operates as an autonomous agent manager. This manager dynamically generates, fine-tunes, and deploys bespoke pricing sub-models tailored for specific asset classes.

This ability to provide an accurate, warrantied value optimised for liquidation scenarios helps lenders in several ways. It gives them the confidence to offer less defensive loan terms, explore new asset types, and drive down their cost of capital, potentially addressing individual needs related to Basel III capital requirements and overall capital efficiency.

Galbraith sees the technology being adopted quickly for hard-to-price assets, where the opportunity for lenders is greatest, whether through offering more competitive terms or simply gaining more confidence in pricing. He stressed that the value proposition of transferring asset price risk is relevant across all sectors of asset-backed lending, including work with both recourse and non-recourse banks and lenders.

Looking at the future of finance, Galbraith believes the role of AI-driven enforceable valuations will reduce risk and move it into “tradeable territory”. He concluded, “Areas, like ours, where margins of error were too high and risk was illiquid, will become more stable, and growth will accelerate because of it.”.