Loan Origination Agent · Use case
How can a bank use an AI agent to prepare loan applications for underwriters on its own infrastructure?
A loan origination agent collects the documents for a loan application, extracts the income and financial figures, and checks them against your written lending criteria. It prepares a file that shows which criteria are met, which are not and what is missing. An underwriter makes the credit decision. The agent runs on your own infrastructure, so applicants' financial records stay inside the bank.
Statements, payslips, forms.
Read one page at a time.
- ✗ 01Underwriters spend much of their time collecting documents and copying figures before they can assess anything.
- ✗ 02Applications hold bank statements and payslips, which cannot be sent to a public AI service.
Where it stops for a person
Always, before a credit decision.Always. The agent gives no approval, decline or price. An underwriter decides, and can see the source document for every figure.
It assembles the application.
An underwriter decides.
Step · 01
The lending criteria are connected from SharePoint or Confluence.
Step · 02
The applicant's documents arrive through File Upload.
Step · 03
The agent extracts the income, debts and other figures and notes where each came from.
Step · 04
It checks each criterion and lists what is met, not met or missing.
Step · 05
An underwriter reviews the file and makes the decision.
Reads from · connectors
File Upload
SharePoint
Confluence
Application documents arrive through File Upload and are read with Markitdown. The lending criteria are read from SharePoint or Confluence.
Acts through · MCP servers
Markitdown
Nothing. It prepares the file for the underwriter.
What you set up
- Connect your written lending criteria.
- List the documents each loan type needs.
- Write test applications with known outcomes.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Does the agent approve or decline a loan?
No. It checks the application against your criteria and an underwriter decides.
Can the underwriter see where a figure came from?
Yes. Each extracted figure points to the document and page it was read from.
Where are applicants' documents processed?
On your own infrastructure, with a model you serve.