Baker Hill

The Clearwater shape at its most literal: a four decade old loan origination system with inference added to it. Strip the models and the origination workflow, the credit analysis, the portfolio risk management and the deposit account opening all remain, and those are the product hundreds of institutions bought. Two details keep this firmly at the lower end of C rather than near the boundary.

The company describes its architecture as AI ready, which is a statement about readiness rather than about a shipped capability. And no separately named generative product has been publicly documented, unlike the two largest platforms in this lane which both ship one under a brand. What is real and does count is automated document extraction, real time data enrichment and pre approval decisioning inside the platform.

An oversight claim is made and it is more than boilerplate, but it is not described. The company states that the platform serves credit risk insights at speed while workflows still govern how the technology operates, so that it works the way banks and credit unions require. That asserts a gate exists without saying what it is, which is squarely the B position on this ladder.

Held off A on the failure this index finds again and again in lending: the successor platform is marketed on instant decisions and pre approval decisioning, and nothing published describes the decline path. No threshold, no exception route, no statement of what a person reviews before an applicant is screened out. In a credit workflow the automated decline is where oversight matters most and it is routinely the one point left undescribed.

Nothing published on accuracy, validation, benchmarking, monitoring or drift for the extraction, enrichment or decisioning capabilities. The absence is pointed in this lane specifically, because the buyers are examined institutions and supervisory expectations for artificial intelligence in credit are converging on explainability and citation of the source behind every figure. A vendor selling decisioning into that environment publishes neither. Recorded alongside the centrality note: an architecture described as AI ready is not a model that has been validated.

The bar met squarely, which is uncommon on this axis. A named client, Mechanics Bank, is published with a quantified outcome, a 200 percent increase in loan production since adopting the origination system, alongside a second named client cited for outcomes and a client success library. Early adopters of the successor platform are named, including a large regional bank that upgraded in February 2026 and a California institution partnered through a dedicated innovation centre.

Operational scale is specific and consistent at more than 7 billion dollars of originations processed monthly. One honest caveat that does not change the grade but should be recorded: the quantified outcome is attributed to the origination system rather than to the inference layer specifically, so it evidences the platform rather than the artificial intelligence.

Nothing published on whether client lending data informs anything the vendor builds, whether tenants are isolated, or what happens to data at termination. The question has real weight for a platform of this shape, since a single vendor holds the credit files of hundreds of competing community institutions in one estate.

No privacy posture, processing terms, sub processor list or retention statement located, for a platform holding loan applications, guarantor financial statements and tax returns for consumers and small business owners across hundreds of institutions.

No certification, attestation or trust portal located in the vendor's own swept material. The phrase built for security and compliance appears as a value statement with nothing enumerated behind it. Banked check with a strong reason to expect a result: a hosted lending platform processing originations for hundreds of supervised depository institutions will have been through their vendor management reviews repeatedly and will hold current reports.

No licence, no supervised test of the product and no programme enrolment located. Flagged as a live open question rather than a settled C, and it is the same one raised on the asset management technology platform built earlier in this session: a firm that operates a hosted service performing lending origination for supervised depository institutions sits on the performing side of the bank service provider line, which is the distinction that moved a core processor from C to B earlier in this sweep. Nothing was located either way and it would move this grade.

The largest exposure on this record by a wide margin, and the second instance of the identical gap in this lane. This is a lending platform running automated document extraction, data enrichment and pre approval decisioning across consumer, small business and commercial credit at hundreds of United States depository institutions.

Fair lending is the one area of financial services where supervisors have decades of enforcement history and where disparate impact is an established doctrine rather than an emerging concern. Nothing is published on fairness testing, protected class analysis, approval rate distribution or independent evaluation. A commercial lending platform built earlier in this sweep carries exactly the same silence, which makes it a pattern in the lane rather than an oversight by one vendor.

No published allocation of responsibility and no route for an affected party, who here is a loan applicant declined or repriced by a workflow this vendor supplies. The institution carries the statutory adverse action obligation, which is the correct legal allocation and also the reason the vendor has never had to publish a position of its own.

No model, provider, architecture or version named for any of the capabilities described. The named cloud document database and analytics fabric are data infrastructure rather than model suppliers, and the distinction matters because it is easy to read one as the other: naming the database your platform runs on tells a buyer nothing about what is doing the inference. Same shape as the deployment disclosure that reads like a supply chain disclosure, recorded elsewhere in this session.

Graded on the principle this axis should follow consistently: A when the vendor is the system of record for the function, B when it integrates with one. This vendor is the lending system of record for hundreds of institutions, carrying origination, credit analysis, risk rating and portfolio risk management, and the successor platform extends that into deposit account opening so it now spans two core banking functions rather than one.

The rebuild names its data infrastructure specifically rather than gesturing at the cloud, and the company states the platform is designed to integrate with each client's existing core capabilities. Recorded as the one thing held back: no core banking provider is named as a certified connector, so the integration with the core itself is described by intent rather than by counterparty.

The hosting stack is named more specifically than most vendors manage, with the successor platform built on a named cloud document database and a named cloud analytics fabric, and the architecture described as cloud first. That is recorded here and not credited, because naming the infrastructure answers what it runs on and not where a client's data sits. No region, tenancy, residency or subprocessing detail was located.

No price, tier, unit or pricing basis published. A competitor's comparison material asserts that enterprise platforms in this lane charge six figures annually with multi month implementations, which is a rival's characterisation rather than a disclosure and earns nothing under the source test.

Hundreds of banks and credit unions with more than 7 billion dollars of originations processed monthly, spanning commercial, small business and consumer lending plus deposit account opening, and reaching from small community institutions up to a regional bank of substantial size.

Held off A on two counts: this is a single jurisdiction operation covering the United States only, and the buyer is one institution type, the depository, with no insurer, asset manager, fintech or non bank lender segment addressed.