AI Cuts Some Deutsche Bank Loan Decisions to Minutes. What Changes?

Credit decisions used to be measured in weeks. For some businesses, Deutsche Bank says AI is reducing that process to minutes. This is what artificial intelligence looks like when it leaves the demonstration stage and enters financial operations.

In an October 26 Reuters interview, the bank’s technology chief, Marie-Jeanne Deverdun, described a project launched during the summer. Certain less-complex corporate lending decisions that previously took weeks can now be completed in minutes. Some multinational lending processes can shrink from months to weeks. Those are specific use cases, not a promise of instant approval for every applicant.

Time is money. In lending, so is judgment.

A company applying for credit may spend days assembling documents while the bank verifies records, checks exposures, tests repayment capacity and coordinates decisions. Shortening the administrative work can reduce friction for borrowers and costs for lenders. A business that receives a decision sooner may be able to purchase inventory, fund a project or address a cash-flow problem before an opportunity disappears.

But speed is not the same as affordability. A quick approval with expensive fees or restrictive conditions can cost more than a slower, clearer offer. An unexplained rejection can be equally frustrating if an automated system misreads unusual business circumstances.

What sits behind the fast answer?

AI can help classify documents, spot missing information and support risk assessment. The bank still needs valid data, documented lending policies and a clear line of responsibility. A system that makes processing faster is not automatically a system that makes credit safer.

Deutsche Bank says cybersecurity is a top priority. It has built an internal AI platform called ADA to monitor and control automated agents and is preparing for a future in which bank agents interact with clients’ own agents. That could make banking more efficient, but it raises questions about authentication, approval limits, confidential documents and who is liable when an automated action goes wrong.

What about banking jobs?

Automation can change work before it eliminates a complete role. Employees may spend less time moving documents and more time handling exceptions, client relationships and complex credit cases. Model oversight, auditing and cybersecurity also require people and expertise. This announcement does not establish how many positions, if any, have been lost because of this lending project.

The lender plans more than €2.5 billion in technology spending from 2026 to 2028, according to Reuters. The investment is substantial; the financial return will depend on measured efficiency, lending outcomes, security and customer trust.

How a small business should respond

Before accepting an offer, compare the total repayment amount, annualized interest cost, variable-rate provisions, fees, collateral requirements and what happens if revenue slows. Ask whether a person can review a disputed decision. Keep records of the documents sent and verify any digital banking request through an official channel.

Money no Sofá: Faster money is not always cheaper money. The best loan is one a business understands and can repay without betting its future on perfect conditions.

Ask NOVA

“How should I compare two business loans when one bank promises a near-instant decision?”

Money Tool: Compare borrowing costs and repayment scenarios.

Read next: The AI boom has a balance sheet.