Artificial intelligence seems unlikely to completely replace enterprise application software like Workday, which companies rely on to store core employee information and financial transaction data. However, AI is imposing potential constraints on the growth of such software.

Some Workday customers are already using AI agents from Anthropic, Microsoft and other vendors to pull data directly from Workday for analyzing employee profiles or business trends. This means customers no longer need to log into Workday's native system.

Five consultants and partners, collectively serving more than 2,000 Workday customers, said in interviews that many enterprises paying for Workday's built-in AI tools, which are layered on top of its core HR and financial databases, are not using these features on a regular basis. Some consultants believe this situation could make it difficult for Workday to charge customers for AI functionality in the future.

IT software company NinjaOne has more than 2,000 employees. Joel Carusone, the company's senior vice president of data and AI, said NinjaOne prioritizes using Anthropic models to build AI agents that can pull data from traditional software systems like Workday and Salesforce to generate HR and sales-related reports. Carusone is not opposed to testing AI products developed by software vendors like Workday, but NinjaOne's selection criteria are very careful. "We don't necessarily have to use the in-app AI that comes with our partner vendors, because we can get all the data we need through our own AI agents," he said.

Workday cites multiple enterprise case studies, saying its AI tools have practical value; company executives say that rising AI tool usage will eventually translate into revenue.

However, Workday's AI deployment has gotten off to a relatively slow start, mirroring the difficulties encountered by many traditional software companies like Salesforce as they try to use AI to drive growth.

Earlier market fears that AI would prompt customers to completely abandon traditional SaaS software, the so-called SaaSpocalypse, have now subsided. But such applications still face a test: they must prove they won't become what customers see as dumb infrastructure, treated merely as a database or system of record, responsible only for storing critical business information that enterprises trust and need to keep accurate. "Primary systems of record won't be skipped. Payroll processing, filing and audit processes still have to run," said Rafael Dornano, a strategic consultant to investment institutions and software companies. This is also why in recent months, Workday's stock price, like Salesforce's, has recovered to some extent. But she believes the risk has shifted to the layer above the system of record, namely the business analytics work that enterprises conduct based on data.

Dornano has her own analytical framework for evaluating software companies in the AI era. Under this framework, Workday will either grow into an AI sales giant, or gradually lose business value as customers use AI to read data within the system, with both possibilities roughly equally likely. "The key question: will humans log into your application, or will AI agents access it?" she said.

Many enterprises are also reluctant to purchase AI services from traditional software vendors like Workday, partly due to rising costs; on the other hand, some enterprises are cutting software budgets and redirecting funds to emerging model providers like Anthropic and OpenAI. Consultants say Workday and other traditional software companies are increasingly adopting usage-based pricing for new AI features rather than bundling them into subscription fees, driving up customer costs. Interviewees mentioned that many enterprises are cautious about using AI in scenarios involving sensitive business data such as HR and finance, fearing that errors could bring costly losses.

To attract customers to use its own AI, Workday has introduced incentive policies: offering free credits for troubleshooting payroll issues and organizing audit materials. Some large customers can use the Sana Enterprise product free for one year. Sana Enterprise is a product launched by Workday that uses AI agents to automate HR, finance and IT workflows, and can also integrate with other business systems customers have already purchased, such as Jira, Salesforce and ServiceNow.

"Our core goal is to increase usage, usage, usage," said Gabe Monroy, Workday's chief technology officer. "As long as we keep focusing on this, I have no doubt monetization will follow naturally. But monetization must be built on excellent products, and that is our top priority right now."

An executive at a consulting firm serving more than 400 Workday customers in North America revealed that about half of the clients have signed new agreements that include usage-based AI features; but only 10% to 20% of customers are actually using Workday AI, such as recruiting agents that help HR teams search for candidates and schedule interviews. Some customers who activated AI features found their bills far exceeded expectations and have asked consultants to help cut costs. Among these 400-plus customers, none are using Sana Enterprise. This is the flagship AI product Workday is betting on, which only launched in March this year. The problem is that Microsoft, Salesforce, ServiceNow and other traditional software vendors have all launched similar products.

Security software vendor Druva, like NinjaOne, has chosen to build its own AI tools to read data within Workday and generate reports related to employee performance and recruiting. Druva CEO Jaspreet Singh said, "We find the built-in agents on the Workday platform of limited use." He stated that Druva spends more than $10 million annually on software and has already purchased Snowflake's expensive AI tools. "If we have to switch back and forth between multiple platforms, each with its own separate agent, that's very inefficient for us."

Setting Up Data "Toll Booths"

Even if customers use their own AI agents to access Workday data, Workday still plans to get a cut. Workday says that, similar to Salesforce and other systems of record, if customers use external AI vendors like Anthropic or Microsoft to read or modify data stored in Workday, Workday will charge a fee. But whether such data "tolls" can become a major revenue source remains difficult to judge.

Dornano cited Microsoft Copilot as an example to illustrate the challenges Workday faces. Customers can use Microsoft Copilot to read HR data from Workday without logging into Workday's native system or using Workday's own AI tools. Although Workday will eventually charge for such data access, it currently lacks pricing power: from May this year to February next year, Workday is waiving overage fees. This means enterprises can access Workday through API interfaces, such as Microsoft Copilot, exceeding subscription plan limits without incurring overage charges. Dornano described this scenario: "Customers are calling Workday, not logging into Workday; and this toll is temporarily waived."

Workday says the grace period is intended to give customers time to plan for future charges. Although the SaaSpocalypse panic that "AI will make customers completely abandon traditional software" has cooled, such software still needs to prove it won't become dumb infrastructure used only for storing data.

Workday has a wealth of positive customer case studies: software company FlexGen says it used Workday AI to review contracts and identify import tax risks, saving $35 million in tariffs; Chipotle used Workday AI to cut hiring time by 75%; water park chain Great Wolf Lodge says it used Workday AI to automate 90% of its hiring process, saving $700,000 annually. In the July fiscal quarter, customers using Workday's own AI agents grew 35% quarter-over-quarter to 5,500; Workday has about 11,500 total customers. The company did not specify how many of those customers are paying for agents versus using free credits or subscription-included allowances.

Workday said in August that its AI agent products and Sana are expected to reach $600 million in annualized revenue, tripling year-over-year. But the software company's overall growth continues to slow, with only 12.8% growth in the July fiscal quarter. That growth rate is not particularly high, only about 2 percentage points above Salesforce, whose revenue is four times larger. According to Koyfin data, Wall Street analysts expect Workday's revenue growth to continue declining in the coming quarters, dropping to 9.7% in the fiscal quarter ending January 2027.

Reuters reported in August that private equity firm Silver Lake had considered taking Workday private. The news energized investors mired in the stock's slump. If the deal materializes, Workday, with a market value of $45 billion, could accelerate its AI product business development away from public market pressures. CEO Aneel Bhusri has remained silent on the privatization rumors. In a recent letter to employees, he neither confirmed nor denied the reports, saying only: "We do not comment on rumors and speculation, and we will not make an exception going forward."

Voting Rights

Workday was founded in 2005 by David Duffield and Bhusri. Together they hold 69% of the company's voting rights. They say this ownership structure stems from Oracle's hostile takeover of PeopleSoft, where the two previously worked, a process that was extremely fierce.

Workday CEO Aneel Bhusri (left) photographed with Stephen Curry in August. Workday went public in 2012 with a highly successful IPO, marking the beginning of a golden era for enterprise software in the cloud age. Today, Workday's HR software is deeply embedded in large enterprises, government agencies, universities and healthcare systems; but customers have long complained about its clunky and difficult-to-use interface. Even Anthropic and OpenAI themselves are Workday customers. Competitors include giants like SAP and Oracle, as well as smaller players like Rippling, Dayforce and Paylocity.

Workday has long charged customers based on employee headcount. But like its peers, last year it introduced usage-based pricing for AI products, and acquired teams and recruited talent to deploy products like Sana. It recently poached several executives from Google: CTO Monroy, Chief Marketing Officer Sarah Kennedy Ellis, and Senior Vice President of Data Cloud Yasmin Ahmed.

Sensitive Data

Ahmed previously led the AI data management software team at Google. She explained her reason for joining Workday: the opportunity to build AI for heavily regulated industries, where enterprises are wary of mainstream large model vendors. "No one wants Google Gemini Enterprise, Anthropic or OpenAI to get their business data," Ahmed said. (An OpenAI spokesperson responded: the company is rolling out new solutions so that when customers use its models, OpenAI cannot see customer inputs or model outputs.)

But the real challenge is that many of these customers are resistant to any AI application in sensitive HR and financial scenarios. "People are full of anxiety about AI plus sensitive data," said Phil Pearson, vice president at consulting firm Kainos. Kainos serves more than 400 Workday customers, and he described these customers as conservative in their approach to AI.

Some competitors and consultants believe Workday has been slow to launch AI features, which may stem from historical technical baggage: the company's software is built on its proprietary programming language XpressO, which current AI coding tools struggle to adapt to. Monroy, who joined last year as CTO, is leading the overhaul to make XpressO more compatible with next-generation AI tools, making it easier for engineers to develop new features.

"The Powerful Energy of Founders Returning"

"The system I inherited was not natively designed to directly accommodate the next generation of generative AI," he said. But now, "Workday's pace of R&D iteration will surprise the outside world."

Many software industry employees are currently worried about layoffs and falling stock prices. Monroy mentioned that Bhusri's return as CEO in February this year, after a two-year pause, boosted employee morale. Workday laid off about 400 people in February; late last month it announced another 500 layoffs, accounting for 2.5% of its workforce. "The energy from the founder's return is real," Monroy said. Bhusri has begun focusing on teams that were previously overlooked. For example, he had originally planned to attend an all-hands meeting of the infrastructure team and stay for 20 minutes, but ended up staying a full hour, answering employee questions about AI strategy.