Custom Loan Origination and Scoring Software for a Bank: 2026 Cost

The verdict in three sentences

When underwriting an SME loan takes three weeks, the best-rated businesses sign elsewhere before your committee even meets. A custom loan origination platform that automates document collection, credit scoring and the decision workflow brings that time down to 5 working days without loosening risk policy. The 2026 ballpark is 150,000 to 350,000 EUR excluding VAT over 8 to 12 months, with an audit trail that satisfies the supervisor.

Why three weeks for an 80,000 EUR loan?

The chief risk officer of a regional bank sees the same bottlenecks every month. The relationship manager chases the client for a missing tax return, re-keys the accounts into an analysis spreadsheet, emails the file to the risk analyst, who sends it back for one more document, then everyone waits for the weekly committee. The file spends 80 % of its life waiting and 20 % being analysed.

The EBA Guidelines on loan origination and monitoring, enforced by national supervisors across the EU and widely used as a benchmark elsewhere, also require each step to be documented: data used, rating model, overrides, decision. An email-based process makes that traceability expensive to rebuild during an inspection.

Custom build or banking package

Off-the-shelf origination packages handle standardised retail lending well. For SME credit, where each bank has its own scorecard, delegation rules and data sources, customising a package often ends up costing as much as a build, with vendor lock-in on top.

A custom build makes sense when the bank processes several hundred SME files a year and wants to keep control of its risk model. Below that volume, a well-chosen package remains a reasonable option.

What the platform contains

The foundation is a client portal where the business uploads documents and tracks its application, a scoring engine whose rules and weights are versioned, a decision workflow with delegated authority by amount and risk level, and a committee module that stores votes, comments and conditions. Every score is explainable: the platform shows which variables weighed in, which meets model transparency expectations.

Integrations often take 25 to 30 % of the budget: core banking, credit bureau, customer master data, document management, e-signature. Security follows sector standards: strong authentication, encryption, separated environments and penetration testing before production.

Kolonell prices this kind of platform on a quote basis, after scoping with the risk department and IT.

Mini case study

Sarah, chief risk officer of a regional bank, underwrites 900 SME loans a year with an average size of 80,000 EUR. The bank estimates it loses 12 % of well-rated applications because of the delay, or 108 loans. With a 2.5 % net interest margin over an average 5-year term, each loan earns about 10,000 EUR. Winning back only half of them, 54 loans, means 540,000 EUR of margin over the life of the loans. On productivity, saving 3 days of analysis on each of 900 files frees 2,700 person-days, more than 4 analysts a year. For a platform at 260,000 EUR excluding VAT, payback comes in under two financial years.

FAQ

How much does a custom loan origination platform cost?

The 2026 ballpark is 150,000 to 350,000 EUR excluding VAT, depending on the number of loan products and integrations. A first phase focused on SME equipment loans sits at the low end.

Does automated scoring replace the credit committee?

No. It prepares the decision and applies delegated authority: small, well-rated files are approved at branch level, the rest go to committee with a complete file. In practice 40 to 60 % of files can fall under delegation.

How do supervisors view an in-house scoring model?

They expect a model that is documented, validated, monitored over time and explainable. The platform versions every scorecard and keeps the detail of every calculation, which eases internal validation and inspections.

How long until production?

Plan 8 to 12 months. A first product can run as a pilot in 2 or 3 branches after 5 to 6 months.

Can we use our historical data to calibrate the score?

Yes, if you have 3 to 5 years of history with defaults flagged. Migrating and cleaning that data often accounts for about 10 % of the budget.

Let's scope your project. One session with your risk department is enough to pick the first loan product, the priority integrations and the pilot timeline. Detailed quote within 48 h. WhatsApp +221 77 596 93 33.