Artificial intelligence is becoming a major investment area for banks, with agent-based systems offering the potential for faster operations and significant cost savings. A new report by McKinsey estimates that banks could recover their investment in AI agents in an average of 10 months, with the use cases studied delivering a projected 3-year return of 2.9 times the initial investment.
The findings are based on an analysis of 52 agentic AI use cases. However, spending on AI tokens, computing and operations could reach 20% of labour costs, making cost control and measurable returns essential.
Banks report potential savings from AI agents
Banks are deploying AI agents across software development, risk management, compliance, operations and customer service. Examples cited in the report include:
- A $20 million investment in software development agents, with an expected 2x return after an 18-month rollout.
- A $30 million investment in mainframe code modernisation, expected to save about $25 million annually.
- A $32 million investment in automated controls, with expected annual savings of $50 million.
- A $5 million investment in customer-support agents, expected to generate 2.8x the initial investment within 1 year.
These are projected returns reported by the banks, not independently verified results.
Customers are increasingly using AI for banking
A survey of 46,000 retail banking customers across 13 countries found that 23% use generative AI for banking-related tasks, while 58% use it for other purposes. Awareness of generative AI has reached 92%.
Among customers using AI for banking, 74% rely on platforms such as ChatGPT, Claude or Gemini, roughly twice the share using their bank’s chatbot. However, 62% would prefer AI-based services and advice from their primary bank.
Privacy remains a concern for 37% of respondents. Meanwhile, 72% of AI users said they would allow an agent to perform at least 1 financial task, although most want approval controls and limits on autonomous spending. The $500 mark emerged as a key threshold.
Productivity, costs and security remain challenges
McKinsey found that 80% of respondents reported individual productivity improvements, but only 6% said these gains had significantly improved earnings before interest and taxes. A benchmark of 7 agentic systems showed task times falling by about 60%.
The report also highlighted cybersecurity risks. FBI data showed that reported losses from AI-enabled crime reached about $900 million in 2025, the first year these losses were tracked separately.
McKinsey said banks must connect AI agents to reliable data, redesign complete processes and establish clear permissions and monitoring. Their success will depend on balancing productivity and revenue gains against operating costs, security risks and effective governance.
Also read: Viksit Workforce for a Viksit Bharat
Do Follow: The Mainstream LinkedIn | The Mainstream Facebook | The Mainstream Youtube | The Mainstream Twitter
About us:
The Mainstream is a premier platform delivering the latest updates and informed perspectives across the technology business and cyber landscape. Built on research-driven, thought leadership and original intellectual property, The Mainstream also curates summits & conferences that convene decision makers to explore how technology reshapes industries and leadership. With a growing presence in India and globally across the Middle East, Africa, ASEAN, the USA, the UK and Australia, The Mainstream carries a vision to bring the latest happenings and insights to 8.2 billion people and to place technology at the centre of conversation for leaders navigating the future.