Banks that ramped up AI budgets saw productivity jump fast. The gap is widening.

Survey shows cash buys efficiency

The American Banker's 2026 AI Talent Shift Survey found a clear relationship between how much banks spent on artificial intelligence and the gains they reported in worker productivity. Organisations that increased AI investment by 25% or more over the past year were far more likely to report moderate to significant productivity gains than those that barely budged their budgets.

The numbers are clear. Roughly 60% of respondents whose organisations upped AI spending by at least a quarter reported meaningful increases in productivity. For firms that raised AI budgets by 10% to 24%, about 40% reported notable improvements. Smaller increases produced far fewer reports of uplift.

The survey also mapped broader workforce impact. Some 64% of respondents said AI had a moderate to significant impact on their organisation’s workforce. Meanwhile, 54% judged that impact as positive.

But 36% said it was still too early to tell, 22% said there was no impact, 3% said AI’s impact was negative and another 3% were unsure.

Those figures reveal two things at once: AI is already changing how work gets done, and adoption is still uneven — both across firms and across roles.

Where the gains are coming from

Respondents were asked what kinds of effects they were seeing. Improved efficiency topped the list, named by 28% of bankers.

Role augmentation — where employees keep their jobs but get AI tools to do them faster or better — came in at 12%. Enterprise-wide deployments and back-office automation were mentioned by 8% and 5% respectively. Workforce reductions were cited by only 3% and decreased morale by 1%.

The headline numbers don’t show how focused the gains really are. The banks spending more are the ones reporting the bulk of productivity wins. That makes sense: deploying AI at scale tends to require money — for talent, for data work, for integrating systems, and for governance. Spend less, and you collect fewer of the benefits.

Tensions underneath the figures

There’s a political economy to the shift. The survey notes firms such as Block and Bolt have announced job losses tied to AI. At the same time, many bankers expect new AI-focused roles to emerge. The two trends can coexist — clerical roles shrinking as new technical and oversight roles appear — but the transition will be messy for many workers.

Mastheads

Surjit Chana, board member of Beneficial State Bank, a Harvard fellow and tech committee member at the Global Alliance for Banking on Values, warned that vulnerability isn’t spread evenly. "The workers most exposed to displacement are those in routine administrative and clerical roles who often lack the financial buffers, professional networks and transferable skills to navigate the transition," he said. "The workers least exposed can be the ones who least need the protection if the transition isn't actively managed with genuine investment in reskilling, honest communication and workforce impact assessments."

Chana’s point is blunt and practical: savings from automation can be real, but the human cost falls unevenly. And without real reskilling, those displaced may struggle to move into the new jobs being created.

Why efficient firms will look like laggards

The impact on comparisons happens quickly and clearly. A firm that spends heavily on AI now will show measurable productivity improvements sooner. Its reports will look strong in quarterly metrics and internal dashboards. Other firms that haven’t spent as much — either because they’re cautious, lack capital, or face regulatory or legacy constraints — will appear to be underperforming in comparison.

It doesn’t mean those firms are worse; it means the metric favors those who invest early and heavily. The result? Benchmarks move quickly, and the league table reshuffles fast. Firms that delay could be penalised by investors, clients and counterparties who read productivity stats at face value.

And that dynamic creates a feedback loop. Stronger-looking firms attract talent and capital, helping them invest more and widen the lead. Slower adopters lose both, making catch-up harder and costlier.

Measurement, governance and real costs

Productivity numbers can be noisy. Banks measure output in different ways. Some savings show up as headcount reductions, others as faster processing or better client service. Not every improvement converts to profit immediately.

There are also governance costs that don’t appear on simple productivity charts: building secure data pipelines, auditing models, complying with regulation, and buying or training staff. Those things take time and money. Richer organisations can absorb those costs faster. Smaller banks may find the ramp-up steep and expensive.

Even though the survey shows only 3% reporting workforce cuts, we should be careful interpreting that number. It captures how many respondents identified job cuts as a primary impact, not the eventual downstream churn across sectors and supplier firms reliant on bank operations. In short, the immediate headline on layoffs understates the scale of change that could unfold over years.

What firms are doing — and what they’re not

Many organisations are combining automation with role redesign, according to the survey. Banks reported role augmentations where staff keep responsibility but work with AI assistance. Others are piloting enterprise-wide projects. Several respondents said they were still assessing impacts.

But not everyone is planning systematic reskilling. Chana stressed the need for "genuine investment in reskilling, honest communication and workforce impact assessments." That’s a plan, not a slogan. It means budgets for training, time to learn new tools, and clear pathways to new roles. Firms that skip that work risk destabilising staff morale and worsening talent shortages in fields they do want to hire for.

What this means for the sector

The immediate takeaway is straightforward: money matters. Organisations that spend on AI — and spend wisely — are already seeing productivity gains. That will change how peers look in simple performance comparisons. It will also force boards and regulators to think about labour transitions and model risk.

But the broader story is less tidy. Gains will be uneven. Some workers will benefit from upgraded jobs. Others will be squeezed. Some banks will surge ahead and attract resources. Others will fall behind and face hard choices.

Right now, the data show a split between early adopters and the more cautious. Whether that gap narrows or widens depends on how firms invest in systems, people and oversight — not just on how much they spend.

Related Articles

"The workers most exposed to displacement are those in routine administrative and clerical roles," said Surjit Chana, board member of Beneficial State Bank, a Harvard fellow and tech committee member at the Global Alliance for Banking on Values.

This article was created with AI assistance.