RNHB leads early adoption of AI in specialist real-estate lending
RNHB is among the leading specialist real-estate lenders in the Netherlands, and of the first in its segment to embed AI directly into the core origination workflow rather than run it as a separate tool. The capability, implemented with Sea.dev, reads submitted documents, extracts the key information, checks it against other documents, and presents it to RNHB analysts for review and acceptance. The decision on a deal always remains with an RNHB employee.
Speed matters in a competitive market: customers and brokers expect a fast, clear answer, and the document-heavy first steps of a deal was the primary bottleneck.
Our origination team was spending a lot of time gathering information from documents and transferring it into our systems. We saw an opportunity to free up time for higher-value work while improving consistency and maintaining data quality throughout the process.
Raida el Ghalbzouri, Risk Manager at RNHB
Specialist lending is disproportionately document-heavy, with a large variety of cases, for example self-employed applicants or non-standard property structures. Previously, the same information had to be read and re-keyed from documents into internal systems by hand: repetitive work that demanded concentration and carried the risk of keying errors entering the process.
The aim was to reduce that repetitive handling and give the origination team more time for credit judgement, customer context and deal complexity.
Embedded in the workflow, with the people in control
The system was integrated directly into the workflow already used by the origination team, rather than requiring users to move into a separate platform. The AI output appears within the existing origination workflow, and requires little credit analyst training.
The AI system does not make lending decisions. It surfaces extracted information and exceptions for an analyst to review, alongside validation checks on individual documents and across multiple sources to flag inconsistencies. Every lending decision remains with an RNHB employee. Before any workflow went into production, RNHB and Sea.dev grounded the system in RNHB’s historical data and agreed a quality threshold that extraction accuracy had to clear before go-live.
Because the system checks information across multiple sources, it began surfacing inconsistencies that a single-document review would not catch. Every analyst correction or rejection is logged and tracked, giving approval and rejection rates across the workflow. Accuracy is already high from the outset and feedback loops raise it over time where it matters most. As a result, the workflow can transition towards exception review: RNHB analysts spend less time correcting repetitive issues and more time understanding the customer, assessing complex lending scenarios and exercising credit judgement where it adds the most value.
Live in weeks, and a foundation for more
The first version went live in eight weeks, with RNHB and Sea.dev working as a single team to validate requirements, refine the workflow and prepare for go-live at RNHB’s pace.
Value arrived with the first production document. The first live case was a complex multi-unit apartment, and it went straight through the new process with high accuracy from the outset.
RNHB sees the project as a foundation for measurable operational improvement: shorter lead times on document-heavy workflows, less manual handling, fewer input errors and earlier visibility of inconsistencies, with the workflow remaining reviewable, measurable and governed. The broader value lies in combining speed with control, and in directing analyst effort towards where human expertise matters most.
RNHB is undergoing an AI transformation, starting with a specific workflow where the problem was well understood, the value was measurable, and the risk could be managed, as an early, deliberate step in its wider AI journey.
This project shows how we approach AI at RNHB: start with a clear business challenge, create measurable value and build solutions that support our people and customers. Intelligent document processing is an important first step in our broader AI journey.
Stephan de Jongh, Head of AI at RNHB
