Bank of America has poured billions into artificial intelligence. Its own research team admits the technology delivers almost nothing to measured productivity so far. Yet the same analysts forecast gains that could eventually dwarf anything seen from electricity or the internet. The contrast captures the tension running through corporate America right now.
Early this year the bank’s global research unit put the current aggregate productivity boost from AI at just 0.1 percent. That figure barely registers against global economic growth of 3.5 percent. “A small aggregate effect relative to all the excitement around AI,” the report noted. Only 20 percent of workplace tasks can be transformed by today’s models. Of those, just 23 percent prove cost-effective to automate at current prices. Labor accounts for roughly half of costs. Automation saves about 27 percent of labor expense. Multiply the pieces together and the theoretical ceiling sits near 0.66 percent before organizational friction, skills gaps and slow rollout compress it to the observed 0.1 percent.
But costs will fall. Models will improve. When they do, the bank argues, productivity gains could run 10 times larger than current estimates. The upside could add a full percentage point to annual global growth over the next decade, lifting the baseline from 3.5 percent to 4.5 percent. And that marks only the beginning. The report describes a classic J-curve: heavy upfront spending followed by accelerating returns once systems embed deeply into operations.
Inside Bank of America itself the picture already looks brighter. More than 90 percent of its 213,000 employees rely on Erica for Employees, the internal version of the bank’s virtual assistant. That tool alone has slashed calls to the IT service desk by more than 50 percent. Developers using a generative-AI coding assistant report efficiency gains above 20 percent. Software testing that once took hours now finishes 90 percent faster in some cases. Bank of America Newsroom detailed these advances in an April 2025 release.
Hari Gopalkrishnan, who oversees technology and information at the bank, described the shift in concrete terms during a November 2025 briefing. A relationship banker who once managed 15 clients can now handle 50 because AI automates preparation of briefing documents and routine follow-ups. “That’s exactly what we’re seeing in the real world,” he said. Erica itself has handled work equivalent to 11,000 full-time employees through more than three billion client interactions. Reuters reported the comments and investment plans.
The bank directs roughly $4 billion of its $13 billion annual technology budget toward new initiatives, many centered on AI. That commitment continues even as Wall Street debates whether the spending spree represents prudent investment or the latest technology bubble. A separate Fortune analysis in May 2026 captured the bank’s own caution. Current data show scant macroeconomic lift. Yet the potential remains enormous once prices drop and adoption widens. Fortune laid out the arithmetic in detail.
Recent commentary sharpens the uncertainty. Deutsche Bank macro strategist Jim Reid warned in early July that meaningful productivity gains from AI likely sit years away. Enterprises still need time to embed the technology fully. If those gains fail to arrive, already elevated debt levels across economies could become harder to sustain. “In my career I haven’t seen anything like AI in terms of potential for productivity,” Reid told Bloomberg Television. “But I would probably caution that it is going to take a number of years for us to properly embed it into enterprises to really get the benefits of that.” Fortune highlighted the risks just days ago.
Bank of America’s own economic institute offers a more optimistic frame. AI-related capital expenditures in software and computing added as much as 1.3 percentage points to U.S. GDP growth in the second quarter of 2025. Small businesses increased payments to technology services by 6.9 percent year-over-year in September of that year, with manufacturing and construction leading. White-collar sectors such as finance show higher AI exposure and stand to capture stronger productivity lifts even if overall job growth remains only weakly correlated. Bank of America Institute published the findings last October.
Yet the gap between internal wins and broad economic data persists. McKinsey’s Global Banking Annual Review projected AI could deliver up to 20 percent net cost reductions for banks in the short term. Those savings, however, will likely erode as competition passes benefits to customers rather than preserving them as profit. PwC analysts suggested fully embracing AI might improve a bank’s efficiency ratio by 15 percentage points, but only after substantial reorganization of front, middle and back offices. Neither firm expects instant transformation.
Aditya Bhasin, Bank of America’s chief technology and information officer, strikes a confident tone. “AI is having a transformative effect on employee efficiency and operational excellence,” he said in the April 2025 release. “Our use of AI at scale and around the world enables us to further enhance our capabilities, improve employee productivity and client service, and drive business growth.” The bank’s patent portfolio backs the claim. More than 1,200 of its nearly 7,400 patents and applications focus on AI and machine learning.
Still, broader adoption data temper enthusiasm. Only about 3 percent of Bank of America households pay for consumer AI services, though that share rose 38 percent from 2024 averages. Higher-income and younger users lead the trend. The bank projects the U.S. consumer AI market could reach $75 billion annually as subscriptions mature and use cases expand across productivity, search and entertainment.
Economists outside the bank add perspective. The Wharton Budget Model estimates AI will raise productivity and GDP by 1.5 percent by 2035, nearly 3 percent by 2055 and 3.7 percent by 2075. Annual productivity growth peaks around 0.2 percentage points in the early 2030s before tapering. The Bank of Canada’s deputy governor noted emerging signs of small gains but stressed the technology’s effects on jobs, inflation and growth warrant close attention. Dallas Fed researchers suggest a reasonable scenario adds 0.3 percentage point to annual productivity growth over a decade.
Bank of America’s research team acknowledges the lag. Diffusion across organizations takes time. Complementary investments in skills and process redesign often prove as important as the models themselves. Without those changes the productivity J-curve flattens. And if the curve never bends upward sharply, the massive capital outlays on data centers, chips and energy could weigh on corporate balance sheets and sovereign debt for years.
So the paradox remains. Inside one of the world’s largest banks AI already frees developers, streamlines research, prepares client materials and handles routine customer requests. Tens of thousands of employee hours shift from busywork to higher-value client work each year. Erica fields billions of interactions. Coders move faster. Bankers serve more customers.
Yet the macroeconomic needle has barely moved. The 0.1 percent contribution looks trivial next to the excitement and the spending. That disconnect explains why investors oscillate between hype and skepticism. It also explains why Bank of America simultaneously touts internal successes, warns of limited near-term GDP impact, and forecasts eventual gains large enough to reshape growth trajectories.
The coming years will test which view prevails. If costs decline as projected and organizations master the complementary changes, the 10x upside could arrive. If diffusion stalls or reliability issues persist, the current drag on measured productivity may linger. Either outcome carries consequences far beyond any single bank’s earnings. For now the data support both optimism inside the enterprise and caution at the macroeconomic level. Banks, investors and policy makers will watch the next wave of adoption numbers closely.
WebProNews is an iEntry Publication