Baker Hill

Baker Hill's product is a lending system of record with four decades behind it and AI added inside it. Without the AI, origination, spreading, credit analysis, portfolio risk and deposit account opening all still run, and those are what its clients bought.

The AI sits in named features: Intelligent Documents and Data extracts and validates figures from financial documents, Ask BKR answers policy and client questions inside the workspace, UN/FY's instant pre approval reads cash flow and collateral, and Opportunity Insights is still in development. Baker Hill's own framing is that AI is not the strategy, better business lending is.

Baker Hill frames its AI as recommendations rather than black box decisions: AI supports the work and informs decisions, the financial institution remains the decision maker, and controls, traceability and decision ownership stay in place. Ask BKR and Opportunity Insights put answers and a second view of the numbers in front of bankers. UN/FY's instant pre approval goes further, using AI driven cash flow and collateral analysis to trigger real time approvals. Baker Hill publishes no threshold, review step or exception route for that path. On the rest of the workflow, the stated design keeps a banker's judgment at the center of the decision.

Baker Hill publishes nothing on accuracy, validation, benchmarking, monitoring or drift for its document extraction, its pre approval decisioning or the assistants built on them, though its buyers are examined institutions and it sells its AI as fit to put in front of an examiner. Traceability shows how an output was reached rather than whether it is right. What Baker Hill does keep is controls, traceability and decision ownership around each recommendation, so an analyst can follow how an output was produced before acting on it.

Mechanics Bank closed 2022 with $157 million of loan production from nine lenders, against about $50 million from 25 before it reworked its lending process with Baker Hill, an increase of more than 200 percent, and its whole lending process now takes three to four business days. Arvest Bank, at more than $27 billion in assets and a client for nearly 30 years, upgraded to UN/FY in February 2026, and Montecito Bank and Trust's chairman and chief executive is quoted on the platform. The outcomes are attributed to the origination system and the process work around it rather than to the AI, so they evidence the platform.

Client data stays isolated to each institution, and the AI runs in a private Microsoft Azure environment with institution specific data isolation, access controls and governance around how it operates. Baker Hill publishes no retention schedule, no statement on whether client data trains its models and no account of what happens to data at termination. The per institution boundary is still stated plainly, which matters for a vendor holding the credit files of many competing community institutions in one estate.

Loan applications, financial statements and tax returns from consumers and small business owners pass through the platform, and Baker Hill publishes no processing terms, subprocessor list or retention statement for that client data. Its stated protection is separation, with each institution's data isolated from every other client's.

Baker Hill states that it maintains SOC 2 Type II and ISO 27001 certifications and that an independent accredited auditor audits its controls and processes every year. The reports and certificate are not posted for download, and no trust portal lists scope, audit period or subprocessors. Its AI runs in a private Microsoft Azure environment with access controls and data isolated by institution, and a chief information security officer sits on the leadership team.

A hosted service that performs loan origination for banks and credit unions puts Baker Hill on the provider side of bank service provider oversight, and it publishes no license, supervised test of the product, program enrollment or statement on whether it is examined as a bank service provider. Its regulatory work faces the institution's own obligations instead, with portfolio risk and current expected credit loss (CECL) compliance tools for risk leaders.

UN/FY's instant pre approval uses AI driven cash flow and collateral analysis to trigger real time approvals, so Baker Hill's AI reaches the credit decision for small business and commercial borrowers at banks, credit unions and finance companies. Baker Hill publishes no disparate impact testing of that analysis and no method for producing adverse action reasons when an application is declined.

Its stated design is transparent recommendations rather than black box decisions, with the institution as the decision maker, controls and traceability kept in place, and each client's AI running in a private Microsoft Azure environment isolated to that institution.

A loan applicant declined or repriced through a Baker Hill workflow deals with the lending institution, which carries the adverse action obligation, and Baker Hill publishes no allocation of responsibility and no route for that applicant. Baker Hill says it builds AI to be examined, and its recommendations stay inside the controls and decision ownership that let a lender trace and answer for each outcome.

Ask BKR, Intelligent Documents and Data, instant pre approval and Opportunity Insights run on no named model, provider, architecture or version. UN/FY runs on Microsoft Azure with Azure Cosmos DB and Microsoft Fabric as its data layer, which shows where the data lives rather than what does the inference.

Baker Hill is the lending system of record at its clients, carrying origination, spreading, credit analysis, risk rating and portfolio risk management, and UN/FY extends that into deposit account opening. UN/FY pulls customer and business data from the core into the application and fetches tax returns, payroll records and bank statements through integrations with ADP, QuickBooks and TurboTax. Its partner directory names the core processors Fiserv, FIS, Jack Henry, Finastra and Corelation, alongside Salesforce, Alloy, Experian, Equifax, TransUnion, CoreLogic and DocuSign.

The hosting stack is named specifically: UN/FY is built on Microsoft Azure Cosmos DB and Microsoft Fabric. Baker Hill states no region or residency commitment and lists no subprocessors, so where a client's data sits within Azure is unstated. What it does state is the separation between clients: the AI runs in a private Azure environment isolated by institution.

Pricing for UN/FY, its AI products and implementation comes through a consultation request rather than a published price, tier, unit or basis.

Banks and credit unions are the core buyers, from small community institutions to Arvest Bank at more than $27 billion in assets, and specialty finance companies have their own offer for statement spreading, financial analysis and credit summaries. Coverage spans commercial, small business and consumer lending plus deposit account opening, with scored, judgmental and Small Business Administration lending inside the business line. The business is United States only.