Baker Hill Turned a Banker’s Spreadsheet Problem Into a $7 Billion-a-Month Lending Machine

Long before “fintech” became a conference badge, Mark Hill was a bank financial analyst with a repetitive problem. Credit analysis demanded too much manual labor. In 1983, he and Karen Hill started Baker Hill with roughly $50,000 from family and friends and a software answer to that stubborn workflow. The founding insight was almost aggressively practical: a lender should not have to reconstruct the same borrower story every time a tax return, balance sheet or loan request lands on the desk.

Forty-three years later, the nouns have changed more than the job. Paper became PDFs. Desktop programs became cloud subscriptions. “Artificial intelligence” joined “statement spreading” in the product copy. Yet Baker Hill still makes its living in the space between a borrower asking for money and a regulated institution deciding what to do. The company says financial institutions use its platform to process more than $7 billion in loan originations each month.

The unglamorous machine behind a loan

Baker Hill sells a configurable enterprise platform to banks and credit unions. It covers commercial, small-business and consumer origination, along with financial spreading, credit analysis, document collection, decisioning, approval routing, covenant tracking and portfolio monitoring. Put plainly, it replaces a relay race among email, spreadsheets, shared drives and disconnected vendor screens with a controlled path through one system.

That breadth is the point. A community bank can buy a narrow consumer application tool almost anywhere. Commercial credit is messier. One borrower may have several entities, guarantors, properties, covenants and years of statements. Baker Hill’s advantage is its memory of those complications, encoded in templates and configurable workflows, while still serving lower-volume consumer and small-business lending on the same platform.

What failed first: the handoff

Across Baker Hill’s public customer stories, the first failure is rarely a bad credit model. It is the handoff. IncredibleBank described customer information scattered among documents and Excel sheets. Rally Credit Union had diverse data-entry points and manual annual-review reminders. Pioneer Bank’s reviews depended on paperwork that could lag the latest account behavior. TowneBank needed to standardize origination as loan volume grew.

The software response is not mysterious. Capture the information once. Make required fields match policy. Route complete files to the right person. Generate a credit memo appropriate to the deal. After closing, watch for overdrafts, late payments, score changes or missing documents. The cleverness lies in handling exceptions without turning every exception into a developer ticket.

“Speed matters, but clarity matters more.”Brad Benson, United Federal Credit Union

Reported outcomes are concrete, though they belong to individual customers rather than a universal guarantee. Mechanics Bank says loan production rose 200 percent while decisioning time fell 85 percent. Rally Credit Union reports eight hours saved per loan and approvals moving 72 hours faster. Pioneer Bank reports a 30 percent reduction in delinquency after adding proactive portfolio monitoring. These gains have a common shape: fewer duplicate entries, earlier warnings and more consistent rules.

A company repeatedly re-underwritten

Baker Hill has also been through its own sequence of approvals. Experian bought the founder-led company in 2005. A decade later, Experian sold it to The Riverside Company for $100 million, explaining that the niche software suite had limited overlap with its larger decision-analytics strategy. Same asset, different owner, different conclusion.

Riverside combined Baker Hill with financial analytics provider DMA and backed a multiyear rebuild. The result was NextGen, a modern SaaS platform that expanded contract values, added products and reached larger institutions. In 2023, Riverside sold the company to Flexpoint Ford, a private-equity investor focused on financial services and healthcare. The price was not disclosed.

What changed minds? For Experian, strategic focus pushed the business out. For Riverside, the same niche offered a platform to modernize and grow. For customers, the decision usually changes when manual work stops being merely annoying and begins limiting volume, hiring, risk visibility or borrower service. A spreadsheet can be cheap for ten loans and ruinous at a thousand.

NextGen grows up, changes its name

Baker Hill is now moving the NextGen suite under a new name, UN/FY. Launched in late 2025, it combines digital small-business lending, deposit-account opening, relationship tools and AI-assisted analysis around what the company calls Data Pond. The architecture pulls from core banking systems, regulatory data and outside partners, using Microsoft Azure Cosmos DB and Fabric to keep information moving.

The product pitch is a less linear loan process: prefill verified information, extract figures from uploaded documents, identify risks, surface the next action and keep a banker in view of the whole relationship. Baker Hill says AI-driven automation can cut origination costs by as much as 60 percent. That is a company claim, not a law of physics. The more interesting design choice is that UN/FY is meant to augment the relationship manager rather than replace one with a score and a rejection email.

The edges are expanding too. A Salesforce integration keeps client-facing employees in the CRM while Baker Hill handles lending logic. LoanPro extends the journey into servicing and collections. Radius, powered by CREtelligent, adds property intelligence. Lumos contributes predictive small-business credit signals. Participate powers an embedded exchange where lenders can distribute loans, improve liquidity and manage concentration. This is how a workflow product becomes an ecosystem: it owns the central record, then lets specialists meet it there.

What a reader can steal

The Baker Hill playbook is surprisingly portable. Start with the expensive repetition, not the fashionable technology. Map every handoff. Decide which information is authoritative. Turn policy into reusable templates. Preserve an escape hatch for expert judgment. Then measure cycle time, re-entry, exceptions and loss signals before promising transformation.

When it will not work

Baker Hill is a poor fit when a lender wants a tiny standalone tool, lacks an executive owner or cannot spare experienced staff for configuration and testing. It will not repair inaccurate source data by decree. It will not make five credit officers interpret policy the same way unless leadership first resolves the policy. And it cannot deliver a clean borrower experience if core systems, document providers and servicing platforms are treated as someone else’s problem.

Its implementation material quietly admits as much. Baker Hill begins before kickoff with goals, roles, preparatory work, weekly agendas, testing milestones and role-based training. That is not administrative garnish. It is the product’s operating condition. Configurability creates leverage only when someone is accountable for the configuration.

In a market crowded with nCino, Abrigo, MeridianLink, Finastra, Fiserv, FIS and homegrown systems, Baker Hill fits best where lending complexity is real but replacing the entire bank stack is unrealistic. It is less a shiny front door than connective tissue - origination deep enough for commercial credit, broad enough for other lending lines and open enough to coexist with the tools a bank already owns.

That positioning also explains the business model. Baker Hill does not post a one-click monthly price because this is not software a branch manager buys on a credit card. Institutions negotiate subscriptions around modules, volume, integrations and deployment scope, then add implementation, training and advisory work. The model rewards long relationships and expansion across lending lines. It also raises switching costs: once policies, templates, documents and partner connections live in the platform, replacement becomes another institutional project. The fair buyer response is to demand measurable milestones before adding modules. A bank should know its baseline approval time, manual touches and exception rate, then release more budget only when the new workflow improves them. Longevity is reassuring; quantified progress is better.

The company’s durable lesson is almost comically unfuturistic. Good enterprise software remembers where the work gets stuck. Baker Hill began with a financial analyst noticing wasted motion. Its AI-era challenge is to keep that observational discipline while the vocabulary races ahead. If UN/FY succeeds, it will not be because a chatbot learned to say “relationship banking.” It will be because a lender stopped chasing a tax return through an inbox and called the borrower back.