Enterprise AI Is Scaling Fastest Where Businesses Can Measure the Results

For all the attention paid to enterprise artificial intelligence strategy, adoption is not spreading evenly across the organization.

According to findings shared in the August 2026 edition of The Enterprise AI Benchmark Report, a PYMNTS Intelligence original, the most important dividing line may not be industry, budget or even executive enthusiasm. Instead, AI appears to be scaling fastest in functions where companies already have structured data, technical ownership and outcomes that can be measured with relative precision.

Among surveyed firms, broad or embedded deployment of new AI tools is already the norm in data and technology. Ninety-five percent of financial services firms report operating at this deeper level of adoption, alongside 84% of healthcare firms and 81% of media firms. Payments and finance tell a more uneven story. Roughly 9 in 10 financial services firms have scaled new artificial intelligence tools in those functions, compared with 63% of healthcare firms and just 43% of media firms. Among media companies, limited deployment remains the most common stage.

The numbers point to an emerging enterprise AI reality: companies are not necessarily scaling AI wherever the technology seems most promising. They are scaling it where the operating environment makes AI easiest to govern, evaluate and improve.

AI Is Following the Data Infrastructure

Data and technology functions are a natural starting point for enterprise AI because many of the prerequisites for deployment already exist. Among firms that have scaled AI across data and technology, 77% are using it for security monitoring, making that the most common application. Infrastructure optimization and data ingestion and cleansing each stand at 68%, while 63% report scaled use of AI governance tooling.

These are not necessarily the most visible applications of generative AI. They are, however, well suited to machine-assisted decisioning. In each case, companies can compare what happened before AI with what happened after it. That makes deployment easier to defend internally and easier to refine over time.

A similar pattern is emerging in payments and finance. The leading scaled applications include treasury and liquidity management, accounts payable and receivable automation and pricing optimization. These functions share many of the characteristics that have accelerated adoption inside technology departments.

Read the report: The Enterprise AI Payback Curve: Adoption Accelerates as Returns Take Shape

Financial services firms appear particularly well positioned to make that transition. About 9 in 10 have already scaled new artificial intelligence tools in payments and finance, according to PYMNTS Intelligence data.

Healthcare firms are further behind at 63%, while media firms trail at 43%. The gap suggests that deploying AI inside a business function depends not only on what the technology can theoretically do, but also on how mature that function’s underlying systems and processes already are.

Feedback loops matter because they give organizations something pilots frequently lack: a mechanism for deciding whether to expand, modify or abandon a deployment.

Companies have spent the past several years evaluating model capabilities, launching pilots, creating governance committees and giving employees access to generative AI tools. The next phase may depend much more heavily on organizational design. Functions built around fragmented databases, manual approvals or loosely defined performance measures create a difficult environment for AI deployment. Even when a model produces useful output, management may struggle to determine whether the system is trustworthy enough for broader use.

An enterprise might have deeply embedded AI systems operating across cybersecurity, data engineering and treasury while still experimenting with artificial intelligence elsewhere. The apparent contradiction disappears once deployment is viewed at the level of business functions rather than the enterprise as a whole.

The post Enterprise AI Is Scaling Fastest Where Businesses Can Measure the Results appeared first on PYMNTS.com.