How AI is helping Deutsche Bank vaporize its tech backlog

Artificial intelligence has supercharged productivity at Deutsche Bank, shrinking technology project timelines that once spanned years down to just a few months, according to a senior executive.

Speaking at the bank's tech event in Bengaluru, India, Denis Roux, Chief Information Officer for the investment bank division, revealed that tasks previously projected to take two years are now being wrapped up in three to six months. Similarly, internal backlogs that used to stall operations for months are now being cleared in mere weeks, News.Az reports, citing Reuters.

Deutsche Bank relies heavily on its tech workforce in India, where roughly 9,000 employees—making up 45% of its global engineering footprint—handle high-value operations like software development, finance automation, and research and development. The bank is currently deploying AI tools to automate complex tasks, such as extracting and analyzing heavy financial data, and building applications that can map external global events against its portfolio to instantly calculate market exposure.

However, the rapid acceleration of AI adoption comes with a catch: rising computing costs. As major AI providers like OpenAI and Anthropic pivot away from flat subscription models toward usage-based "token" pricing, the bank is introducing strict budget discipline.

To prevent runaway expenses, Deutsche Bank engineers are allocated specific token quotas. If a team needs additional computing capacity, they must formally request it and prove the tangible business value of their project.

Roux noted that the bank remains highly cautious about using AI as a blanket solution for everything. Moving forward, the lender plans to save its most advanced AI capabilities for high-impact projects, opting for simpler models or traditional software solutions for routine everyday tasks to ensure a strong return on investment.

News.Az 

By Aysel Mammadzada