Introduction
Zest AI builds machine-learning software that banks, credit unions, and specialty lenders use to underwrite consumer credit — deciding who qualifies for a loan and on what terms. The company, founded in 2009 and originally known as ZestFinance, has spent much of its history arguing that traditional credit scoring leaves usable information on the table, and that better statistical models can approve more borrowers without taking on more risk.
Unlike consumer-facing lenders, Zest AI sells its underwriting technology to institutions that keep the lending decision, and the regulatory responsibility for it, in-house. That makes the company a useful lens on how AI is adopted inside heavily regulated financial institutions rather than at newer, AI-native startups.
What the company does
Zest AI’s flagship offering, Zest Automated Machine Learning (ZAML), lets lenders build, validate, and monitor machine-learning credit models rather than relying solely on conventional credit-bureau scorecards. The company positions the product around explainability — giving lenders and their examiners a way to understand why a model made a given decision — since underwriting models used by regulated institutions must be able to justify adverse-action notices to declined applicants.
Who it serves
Zest AI sells primarily to banks, credit unions, and specialty lenders across consumer, auto, and small-dollar credit categories, rather than to individual borrowers. Its own company materials describe supporting a mix of institution sizes, including credit unions using its models to expand access to credit for their members.
Company background
Zest AI was founded in 2009, originally as ZestFinance, according to the company’s own “about us” history page, which describes 2009 as “the beginning” of its story. Multiple industry profiles report that the company was founded by Douglas Merrill, a former Google chief information officer, alongside colleagues from the credit industry, though Zest AI’s current official about page does not itself list individual founders, so this profile does not state founder names as confirmed fact. In July 2013, the business — then known as ZestCash — was renamed ZestFinance and raised $20 million in funding that included participation from Peter Thiel, according to contemporaneous reporting. In October 2019, marking its tenth anniversary, the company rebranded from ZestFinance to Zest AI, a change it described as reflecting its evolution into an enterprise software provider focused on automated, explainable machine learning. The company is headquartered in Burbank, California.
Product and AI capabilities
ZAML is built around applying machine-learning models — rather than traditional linear scorecards — to underwriting decisions, using a larger set of variables than conventional credit scoring typically considers. Zest AI emphasizes model governance features such as adverse-action explainability and fair-lending bias testing alongside the underwriting models themselves, positioning the software as a way for regulated lenders to adopt more sophisticated statistical models without losing the ability to explain individual decisions to regulators or declined applicants. According to Zest AI’s current company page, its technology has supported $1 trillion in loan originations and includes more than 600 active models, with more than 50 patents issued or pending.
Key developments
Zest AI traces its founding to 2009, when it launched as ZestFinance. In July 2013, the ZestCash brand was renamed ZestFinance alongside a $20 million funding round that included participation from Peter Thiel. The company rebranded from ZestFinance to Zest AI in October 2019, at its ten-year mark, repositioning itself around the ZAML underwriting platform. Its company timeline also references the creation of a credit union service organization structure in the years that followed, reflecting closer ties to credit union customers. As of its most recent published figures, Zest AI says its models have supported $1 trillion in loan originations.
Why it matters
Zest AI is a useful case study in how AI adoption looks inside regulated lending, where the technology has to satisfy fair-lending law and examiner scrutiny as much as it has to improve accuracy. Its emphasis on explainability, rather than raw predictive power alone, reflects a constraint that many AI-in-finance narratives skip over: a model a bank cannot explain to a regulator is often not a model a bank can deploy at all. That trade-off is a recurring theme across underwriting-focused companies in this sector.
Sector context
Zest AI sits within Brel’s AI in finance coverage as an underwriting-technology vendor, distinct from consumer lenders such as Upstart that use similar techniques to originate loans directly. Readers comparing underwriting approaches may also find it useful to look at identity and fraud-focused companies such as Socure, which address a different part of the same lending pipeline.
Sources and references
This profile is based on Zest AI’s own company history page and independent reporting on its 2019 rebrand and funding history.
- Zest AI — “About Us” company history timeline (zest.ai)
- National Mortgage Professional — “ZestFinance Rebrands as Zest AI” (2019)