Banks should focus their AI spending on strengthening core business areas where they have significant market presence, according to Alexandra Mousavizadeh, co-founder and co-CEO of Evident, a firm that tracks AI adoption in finance. While basic AI applications like document review and onboarding are now essential—what she calls “table stakes”—the real competitive advantage lies in deeply integrating AI into a bank’s most strategic operations. For wealth management-focused institutions, this means enhancing advisors’ ability to analyze client data and deliver personalized insights. For retail banks, it could mean improving customer engagement through smarter chatbots and more intuitive digital experiences. The goal is not just automation, but creating a sustainable edge through AI-driven decision-making and service innovation. AI is no longer optional in banking. According to ThoughtLinks, AI is expected to reshape 44% of banking work by 2030. JPMorgan Chase, currently ranked first in Evident’s AI maturity index, has invested at least $2 billion in AI and is rolling out tools across its 300,000+ workforce. Yet, despite massive spending, many investors and analysts are asking when the returns will materialize. Bank executives are now under pressure to demonstrate tangible improvements in productivity and revenue. Mousavizadeh noted that banks with centralized technology decision-making tend to move faster and achieve better integration. This contrasts with earlier “crowdsourcing” models, where individual teams adopted AI tools independently, often leading to fragmented results. Dan Priest, PwC’s chief AI officer, echoed this, saying top-down approaches yield better ROI by focusing on fewer, high-impact tools and enabling deeper expertise. AI agents are emerging as a key focus, particularly for automating back-office tasks. Goldman Sachs has been collaborating with Anthropic to develop co-autonomous agents that handle trade accounting, transaction processing, and client onboarding. The bank’s tech leadership expects these tools to launch soon, with no immediate plans for widespread job cuts. Measuring AI success has also evolved. Banks are shifting from tracking isolated use cases to scaling capabilities across departments. The focus is now on building internal AI architectures that allow workflows to be reconfigured and reused across lines of business. To achieve this, banks must combine top-down strategy with bottom-up adoption. Mandatory AI training helps ensure widespread access, but Mousavizadeh stressed that culture matters too—banks need to foster creativity and experimentation. Looking ahead, the ultimate goal is to define a vision for a fully AI-integrated bank. “You need to work backward from that future state,” she said. This means designing new systems and processes not just to improve existing products, but to reinvent how banking operates—doing so faster than competitors to stay ahead.