Speakers from McKinsey & Company told attendees at Portfolio's Banking Technology 2026 conference that personal AI agents are changing how retail banking works: they accelerate innovation and time-to-market, enable highly personalized customer offers, but also create new competitive dynamics and mounting cost pressure.
András Havas, partner at McKinsey & Company, highlighted four major trends: the industry is shifting from AI hype to a focus on real value creation; personal AI agents are becoming mainstream; AI sovereignty is gaining importance (with EU initiatives as an example); and AI security incidents are eroding trust.
What AI agents do and why it matters
AI agents first ingest and analyze data, then present optimization opportunities for users' finances. They are increasingly used to reduce information asymmetry: McKinsey reports that one in four British customers already consults AI agents to optimize their finances. The next phase may see agents contacting banks directly on behalf of customers — even negotiating — which raises strategic questions for banks about how to handle third‑party agents. Banks are responding with restrictions, partly for security reasons and partly to protect their own business models.
Economic impact and key figures
McKinsey estimates that interactions mediated by agents could put about $170 billion of banking profit at risk by 2030 — roughly 9% of the sector's total profit pool. A UK consumer survey cited by McKinsey found that 64% of customers most trust their own bank even when considering AI-based suggestions.
On the cost side, Etele Bosits, associate partner at McKinsey & Company, noted two trends: progressing from experimentation to individual use cases can drive costs 2–4 times higher, and many Hungarian financial providers face rising AI expenses that could amount to a quarter of their IT budgets. Although almost every institution now uses AI capabilities at some level, only about one third report positive financial results and only 6% classify as AI 'high performers.'
How banks can realize value
McKinsey's speakers emphasized that capturing AI value requires domain-level transformation rather than isolated projects. Transformations that rewire entire teams and processes can deliver much bigger gains; domain-wide implementations can cut costs by up to 50% compared with approaches that start from single use cases. Most technology leaders expect AI costs to rise, and many institutions have already introduced controls to prevent runaway spending. Successful programs therefore need to build the economics of AI into use-case development from the outset and rethink the banking business model as a whole.
Main takeaways
Customers generally trust financial providers, and that trust must be preserved through demonstrable value creation. Banks will need to rewire their operating models, keep value creation central, and manage risks related to AI sovereignty, security incidents, and agent-mediated interactions if they want to benefit from the new AI-driven landscape.
Key dates and references
- McKinsey presentations at Portfolio's Banking Technology 2026 conference (session took place on a Tuesday).
- Portfolio AI & Digital Transformation conference scheduled for November 26.
- Related Portfolio articles dated 2026-10-22 and 2026-11-10.
Tags: banks, artificial intelligence, digitalization, banking sector, digital transformation, McKinsey, Banking Technology, AI agents
