Deutsche Bank is using artificial intelligence to accelerate parts of its corporate lending process, reducing the time needed for some smaller business loan decisions from weeks to minutes. The development offers a concrete example of how AI is moving beyond customer chatbots and into the operational systems that underpin major financial institutions.
In an interview with Reuters published on 26 October 2026, Marie-Jeanne Deverdun, the bank’s technology chief and a member of its management board, described how the initiative is changing the speed at which certain lending decisions can be made.
The project went live during the summer. According to Deverdun, it can also shorten some decisions involving multinational companies from months to weeks.
The results apply to particular cases, not every corporate loan. Nevertheless, they illustrate the potential for AI to reduce administrative delays in a sector where decisions often depend on extensive documentation, risk analysis and internal approvals.
Why corporate lending takes time
Corporate lending is more complicated than assessing a simple consumer loan.
Banks must evaluate a company’s financial position, cash flow, existing liabilities, business model and ability to repay. Larger borrowers can have complex ownership structures, international subsidiaries, multiple currencies and legal arrangements across several jurisdictions.
Information must be collected, checked and interpreted before a bank can determine whether a loan is appropriate and under what conditions it should be offered.
Much of this work involves structured information, document analysis and repetitive administrative tasks. These are areas where AI systems can potentially assist by organising records, extracting relevant details and supporting employees as they prepare a decision.
The technology does not remove the need for risk management. Instead, it can help employees reach the stage where a decision can be made more quickly.
For smaller companies with comparatively straightforward structures, the potential gains may be particularly noticeable. A process that once involved several hand-offs between teams could become faster if information is available in a consistent format and routine steps can be automated.
A broader transformation of banking
Deutsche Bank’s initiative forms part of a wider effort to make the institution more technologically integrated.
Chief executive Christian Sewing has spoken about the ambition to make Deutsche Bank a fully AI-powered bank. That ambition involves more than installing a chatbot or adding a generative AI feature to an existing application.
Banks operate complex technology estates built over many years. Systems responsible for customer records, payments, risk calculations and compliance may use different architectures and data formats. Connecting them securely is often as important as selecting an AI model.
Deverdun told Reuters that Deutsche Bank has developed an internal AI platform called ADA, named after computing pioneer Ada Lovelace. The platform is designed to help the bank develop, monitor and control AI agents, oversee their activities and manage risk. It can also support the use of different AI models as technology evolves.
This approach reflects a growing distinction between experimentation and enterprise deployment. An AI tool may perform well in a demonstration, but a bank needs systems that can be monitored, audited and governed in a regulated environment.
From AI assistants to AI agents
One of the more significant developments is the bank’s preparation for a future in which its AI agents interact with agents operated by clients.
An AI assistant typically responds to a request or helps a person complete a task. An AI agent can be designed to carry out a sequence of actions towards a defined goal, subject to its permissions and controls.
In a corporate banking context, a client’s agent might eventually help prepare a financing application, organise financial information or request updates. A bank’s agent could process the information, check whether required documents are present and coordinate parts of the assessment.
Such interactions could reduce manual work on both sides. But they also introduce new questions about identity, accountability and authority.
How can a bank verify that a client’s agent is authorised to submit a request? Who is responsible if an automated system sends inaccurate information? What happens if an agent acts on a malicious instruction or accesses data beyond its intended purpose?
These are not minor technical details. They determine whether automated interactions can be trusted in a sector responsible for sensitive financial information.
Cybersecurity remains a central concern
Deverdun told Reuters that managing cyber risk was her top priority. That concern reflects a wider challenge facing banks as they introduce more connected AI systems.
AI agents may need access to documents, internal applications and customer information to complete their tasks. Every additional permission can create another potential route for misuse if the system is compromised or behaves unexpectedly.
A secure deployment therefore requires more than accurate model outputs. Banks need strong access controls, monitoring, clear boundaries on what agents can do and the ability to intervene when something goes wrong.
Regulatory expectations also remain important. Lending decisions can have significant consequences for businesses, and institutions must be able to explain and justify the processes they use. Faster processing should not come at the expense of fair treatment, appropriate risk assessment or legal compliance.
What businesses could gain
For businesses seeking finance, faster decisions could improve planning and reduce uncertainty. Companies often need working capital to purchase equipment, pay suppliers, expand operations or manage temporary cash-flow pressures.
A quicker decision does not guarantee loan approval, lower interest rates or more favourable terms. It can, however, reduce the time spent waiting for a response when the application is straightforward and the necessary information is available.
For banks, automation may improve productivity and allow staff to focus on complex cases that require judgement and specialist expertise.
Deutsche Bank plans to spend more than €2.5 billion on technology between 2026 and 2028, according to Reuters. That commitment shows that AI deployment is part of a broader technology investment programme rather than an isolated experiment.
The bottom line
Deutsche Bank’s use of AI in corporate lending demonstrates how automation can affect a high-value business process. Cutting some decisions from weeks to minutes could make financing workflows more responsive, but the benefits will depend on the quality of the data, the complexity of individual applications and the controls surrounding automated systems.
The bank’s next challenge is to scale these capabilities without weakening security, accountability or the quality of credit decisions.
If it succeeds, AI could become less visible to customers while having a substantial effect on how quickly the financial system operates.
Source: Reuters, 26 October 2026 — AI cuts some Deutsche Bank corporate loan decisions to minutes.
Laat een reactie achter