• AI adoption trends: Around 80% of surveyed banking CXOs expect a major share of future technology budgets to go toward AI, building on scaled use cases like credit underwriting and collections voice bots that operate at one-fourth to one-fifth the cost of human telecallers.
• Projected operational returns: Deploying AI at scale is expected to decrease the cost to serve by at least 30% to 40% while lifting productivity by 50% to 60% for middle- and back-office roles.
• Token cost management: Financial institutions are lowering token-related expenses by 60% to 70% through token caching and request routing between small and large language models, with token costs expected to fall by one-fourth to one-fifth within a year.
• Change management focus: Under BCG's 10-20-70 framework, 70% of the effort to unlock AI value relies on people and process transformation, compared to only 30% for models and technology.