Banks in the United States are continuing their investment in AI, which will have both short- and long-term effects on their efficiency ratios. 

Publicly held banks are expected to experience a slight increase in their average efficiency ratio — from 56.3% in 2025 to 56.9% in 2026 and 57.1% in 2027, according to Deloitte’s 2026 banking and capital markets outlook report published Oct. 30. 

The efficiency ratio is a financial metric that measures how effectively a bank uses its resources to generate revenue or profit, according to the Federal Reserve. A lower efficiency ratio reflects better management of funds to generate revenue. 

According to the Deloitte report, smaller banks outnumber bigger banks and will need more time to realize efficiency gains from AI investment, thus a temporary spike in efficiency ratio is expected, but over time, it is expected to trend down.

However, efficiency ratios for banks with clear AI strategies are likely to improve in 2026, supported by tighter expense management and continued gains from digitalization and AI-driven efficiencies, Val Srinivas, banking and capital markets research leader at the Deloitte Center for Financial Services, told FinAi News. 

“Overall, both large and regional banks are expected to see progress, though the degree of improvement will vary across institutions,” Srinivas said, adding that larger banks are expected to see their efficiency ratios improve quicker than smaller banks due to continued investment in technology. 

Smaller banks must invest not only in AI initiatives but preparing a bank to deploy new tech that will affect their efficiency ratio in the short term, Srinivas said. 

However, some of the biggest banks are already seeing their efficiency ratios improving alongside their tech spend.

JPMorgan, for one, increased its tech and communication spend by 16% year over year in Q3 2025 to $2.8 billion and reported its efficiency ratio drop to 52% in Q3 from 52.4% during the same period last year, according to the bank’s Oct. 14 earnings report. 

(Global Systemically Important Bank Efficiency Ratio)

Citi, similarly, reported Q3 2025 tech and communication spend of $2.3 billion, up 2% YoY, while its efficiency ratio clocked in at 64.7%, an improvement from 65% in Q3 2024, according to the bank’s Oct. 14 earnings report. 

ALSO VIEW: FinAi News’ Efficiency database here. 

Need for cohesive AI strategy 

Many banks lack an overarching vision for AI, causing pilot programs to be sporadic, isolated and poorly tracked against key performance indicators, Srinivas said.  

“One reason for this is a lack of ownership and accountability across the AI lifecycle, in particular, a central entity that oversees a roadmap for enterprisewide execution while meeting the needs of business lines,” he said. “Without this framing, banks may fail to establish concrete outcomes aligned to their broader mission or struggle with undisciplined funding.” 

Without a unified vision, banks may struggle to identify scalable AI opportunities, the report stated. 

Many banks, including Ally Financial, TD Bank and Academy Bank, have created AI centers of excellence to centralize AI-driven ideas and their implementation.  

Finding ROI if a proper AI strategy isn’t in place can be difficult, Srinivas said. 

“There’s no single approach for calculating ROI from generative AI,” he said. “The approach depends on each institution’s goals and context.” 

What matters most is clarity and a disciplined way to assess results, even if they are purely quantitative, Srinivas said.  

Some outcomes will be indirect. Shorter customer service calls, for example, might boost satisfaction and, over time, increase cross-sales, he added.  

Despite not finding a clear ROI number, banks should continue to invest in AI, Srinivas said, adding that some “gains are often hard to measure precisely.” 

Register here for early-bird pricing for the inaugural FinAi Banking Summit 2026, taking place March 2-3 in Denver. View the full event agenda here.