Blab Tech
Deutsche Bank Uses AI to Cut Project Timelines, Cautions on Rising Costs
At a packed session of the Bank on Tech event in Bengaluru, India, Deutsche Bank’s chief information officer for investment banking, Denis Roux, revealed that the bank is shrinking technology project timelines from years to months by harnessing artificial intelligence.
Roux, who spoke on Thursday without disclosing exact figures, explained that tasks that once required two years of effort can now be finished in three to six months. He added that backlogs that used to take months to clear are now being resolved in weeks. The senior executive said the bank’s goal is to use AI tools “to continue to make things more efficient.” These remarks come as Deutsche Bank’s technology workforce in India—about 9,000 people, roughly 45 % of the bank’s global tech staff—shifts toward higher‑value functions such as software development, finance, and research and development.
The bank is building AI applications that automate routine data‑processing tasks, including the extraction and analysis of financial data. It is also developing tools that link external events—such as geopolitical developments or market movements—to its investment portfolio in order to assess exposure. Roux said the bank remains cautious about deploying AI for every use case, preferring simpler models for routine tasks and evaluating where traditional solutions may still be more effective.
A key challenge for Deutsche Bank is the cost of AI services. The bank’s engineers are assigned token quotas that limit the amount of AI usage each can consume. When additional capacity is needed, engineers must demonstrate the value that the extra usage will bring, after which the learning is shared across the organization. Roux compared the discipline required to manage token usage to the discipline the bank developed during its transition to cloud computing. He noted that providers such as Anthropic and OpenAI are moving away from subscription‑based models toward token‑based pricing, charging customers based on the number of tokens processed.
The bank monitors usage patterns to avoid bottlenecks while still ensuring a return on investment. According to Roux, the goal is to keep engineers productive without incurring unnecessary costs. The token‑quota system also allows the bank to track which AI applications deliver the most business value and to adjust spending accordingly.
Deutsche Bank’s approach reflects a broader trend in the financial services industry, where large institutions are adopting generative AI to accelerate internal processes while remaining vigilant about cost and risk. The bank’s focus on India’s tech hub aligns with a global shift toward using Indian resources for higher‑value work.
At present, Deutsche Bank has not announced any new AI‑powered products or regulatory filings related to its AI initiatives. The bank’s current strategy involves continued development of internal AI tools, careful monitoring of token usage, and selective deployment of AI solutions that demonstrate clear efficiency gains.
The bank’s next steps will likely involve evaluating the long‑term return on its AI investments and determining how to scale successful applications while maintaining cost discipline. The industry will watch how Deutsche Bank balances rapid productivity gains with the rising costs of token‑based AI services.
Roux, who spoke on Thursday without disclosing exact figures, explained that tasks that once required two years of effort can now be finished in three to six months. He added that backlogs that used to take months to clear are now being resolved in weeks. The senior executive said the bank’s goal is to use AI tools “to continue to make things more efficient.” These remarks come as Deutsche Bank’s technology workforce in India—about 9,000 people, roughly 45 % of the bank’s global tech staff—shifts toward higher‑value functions such as software development, finance, and research and development.
The bank is building AI applications that automate routine data‑processing tasks, including the extraction and analysis of financial data. It is also developing tools that link external events—such as geopolitical developments or market movements—to its investment portfolio in order to assess exposure. Roux said the bank remains cautious about deploying AI for every use case, preferring simpler models for routine tasks and evaluating where traditional solutions may still be more effective.
A key challenge for Deutsche Bank is the cost of AI services. The bank’s engineers are assigned token quotas that limit the amount of AI usage each can consume. When additional capacity is needed, engineers must demonstrate the value that the extra usage will bring, after which the learning is shared across the organization. Roux compared the discipline required to manage token usage to the discipline the bank developed during its transition to cloud computing. He noted that providers such as Anthropic and OpenAI are moving away from subscription‑based models toward token‑based pricing, charging customers based on the number of tokens processed.
The bank monitors usage patterns to avoid bottlenecks while still ensuring a return on investment. According to Roux, the goal is to keep engineers productive without incurring unnecessary costs. The token‑quota system also allows the bank to track which AI applications deliver the most business value and to adjust spending accordingly.
Deutsche Bank’s approach reflects a broader trend in the financial services industry, where large institutions are adopting generative AI to accelerate internal processes while remaining vigilant about cost and risk. The bank’s focus on India’s tech hub aligns with a global shift toward using Indian resources for higher‑value work.
At present, Deutsche Bank has not announced any new AI‑powered products or regulatory filings related to its AI initiatives. The bank’s current strategy involves continued development of internal AI tools, careful monitoring of token usage, and selective deployment of AI solutions that demonstrate clear efficiency gains.
The bank’s next steps will likely involve evaluating the long‑term return on its AI investments and determining how to scale successful applications while maintaining cost discipline. The industry will watch how Deutsche Bank balances rapid productivity gains with the rising costs of token‑based AI services.
