The annual Workday Rising conference in Las Vegas kicked off this week with enterprise software giant Workday placing its primary strategic bets firmly on autonomous artificial intelligence agents designed to fundamentally re-imagine the nature of corporate work rather than merely answer routine queries. However, even as the vendor champions its new agentic capabilities across global markets, real-world deployment challenges are emerging in the enterprise sector. For manufacturing conglomerate Dow, a critical operational hurdle has temporarily halted the rollout of these advanced AI features: the introduction of consumption-based pricing models, specifically Workday’s Flex Credits framework, which has sparked complex internal negotiations between IT and Human Resources departments regarding budgetary responsibility, long-term accountability, and consumption governance.
Dow currently operates under Workday’s Universal Master Service Agreement, utilizing the cloud platform extensively for front-end HR service delivery and general employee management across numerous international jurisdictions. While the enterprise has successfully integrated a substantial volume of traditional machine learning and standard AI features within its primary Workday tenant, it has yet to activate any autonomous agents. Although corporate leadership remains firmly committed to deploying these capabilities in the near future, the Flex Credit pricing model has introduced a strategic roadblock that requires careful institutional navigation before any operational switch is flipped.
The Shift from Seat Licenses to Consumption Models
The core friction point centers on how enterprise software is purchased, billed, and monitored. Traditional software-as-a-service (SaaS) procurement relies heavily on per-user seat licenses or specific software stock-keeping units (SKUs). Under this historical model, HR departments traditionally collaborated with corporate sourcing and procurement teams to secure fixed budgets for specific technology investments designed to enhance workforce productivity and performance. Once a SKU was purchased and provisioned, usage could scale organically across the designated user base without incurring variable financial penalties or unexpected monthly overages.
The advent of agentic AI upends this conventional procurement architecture. Workday’s Flex Credit model decouples software costs from static seat licenses, shifting the financial burden toward a consumption-based metric that measures the actual volume of work carried out by autonomous digital agents. This fundamental evolution creates an immediate organizational ambiguity: because HR traditionally relies on IT to fund, procure, and technically manage enterprise SaaS providers, determining ownership of consumption governance becomes a delicate internal balancing act.
Sarah Jo (SJ) Rowley, HR Digital Product and AI Leader at Dow, highlighted this exact structural dilemma during discussions at the Las Vegas summit. Rowley noted that while the transition to Flex Credits eliminates traditional funding stage-gates—granting teams flexible access to deploy agents as needed—it simultaneously introduces the unprecedented risk of financial overages. Without a clear internal framework dictating who monitors consumption velocity and manages budget variances, departments face a governance vacuum. IT traditionally pays the vendor invoices and oversees the technical support model, yet HR is the functional beneficiary dictating how the agents are utilized. Consequently, enterprises must establish entirely new cross-functional oversight committees to prevent runaway cloud and AI expenditures.
Workforce Skills Degradation and Affordability Challenges
Beyond budgetary governance, Dow’s cautious approach to agentic AI is heavily influenced by concerns surrounding workforce skills utilization and long-term operational resilience. When an organization delegates routine administrative tasks, compliance checks, or service delivery inquiries to an autonomous digital agent, human employees naturally cease performing those specific duties. Over time, this operational offloading risks degrading the core competencies of the personnel who traditionally managed those workflows.
Rowley points out that the traditional conversation around automation often ignores whether a task remains relevant once an agent takes it over; the immediate consequence is simply the removal of the human element from that specific process loop. If an enterprise achieves high adoption rates for autonomous agents across multiple business units, a secondary affordability challenge emerges: the organization must carefully calculate whether the financial cost of the consumed Flex Credits scales more rapidly than the labor efficiencies gained.
This dynamic introduces a completely new mindset for corporate leadership. When utilizing traditional static SKUs, organizations could absorb fluctuations in user adoption without financial penalty. Under a consumption-based agentic model, however, widespread adoption could inadvertently trigger severe budgetary overages. This reality forces enterprises to designate internal stakeholders responsible for babysitting consumption metrics, establishing contingency plans for budget depletion, and mitigating the legal and operational liabilities tied to automated actions.
The Broader Implications of AI Accountability
As enterprise software vendors accelerate their investments in agentic innovation to recoup significant research and development outlays, corporate customers are forced to redefine the parameters of artificial intelligence governance. Responsible AI use can no longer be confined strictly to data privacy compliance, algorithmic bias mitigation, and security guardrails. In an agentic era, responsible AI inherently encompasses financial affordability, operational accountability, and direct liability.
Even when an autonomous agent executes a task on behalf of an enterprise, the corporate entity retains full legal and operational liability for the outcomes of those actions. Determining where that liability resides within the corporate hierarchy—and which department is tasked with continuously auditing agent behavior—remains an unresolved challenge for many Fortune 500 organizations. These structural ambiguities help explain why major enterprises are deliberately pausing deployments to design robust internal governance models before exposing autonomous workflows to global employee populations.
Regulatory Fragmentation and the Compliance Labyrinth
Compounding internal governance challenges is the increasingly complex external regulatory landscape governing artificial intelligence technologies. In the United States, the absence of comprehensive federal legislation establishing standardized AI guardrails has left a regulatory vacuum that individual state and municipal governments are aggressively filling. This legislative fragmentation creates a formidable compliance burden for multinational and nationwide enterprises operating across multiple jurisdictions.
Rowley emphasized that navigating a patchwork of disparate state and city-level regulations—such as New York City’s localized algorithmic accountability laws—makes maintaining a consistent, friction-free user experience exceptionally difficult. While global enterprises are generally accustomed to adapting regional compliance frameworks for Europe, Asia, and the Americas, navigating hyper-localized municipal mandates introduces a level of administrative friction that could stifle enterprise innovation. Industry experts suggest that without a cohesive, top-down federal framework for AI deployment, multinational firms will be forced to implement complex, region-specific policy filters just to run standardized enterprise software applications safely.
Process Optimization as a Prerequisite for AI Integration
Despite these formidable governance, cost, and regulatory hurdles, Rowley and other enterprise technology leaders are far from AI skeptics. The pace of technological maturation over recent months has been striking, with generative and agentic models demonstrating a sophisticated capacity to understand context, personalize research, and streamline professional workflows. For Dow, the path forward involves recognizing that successful agentic deployment requires rigorous internal process optimization before advanced features are activated.
Dow originally went live on Workday in 2019, embarking on a cloud migration journey that required the manufacturing firm to fundamentally adapt its internal operational processes to remain agile alongside continuous vendor innovations. Operating a modern SaaS environment successfully requires an organizational willingness to adopt new features as they become available; failing to turn on advanced capabilities diminishes the long-term return on investment inherent in modern cloud subscriptions.
While Dow maintains exceptionally high user adoption rates across standard Workday modules—such as personal data management and job transitions, where digital execution is mandatory—many broader workflows still require significant human support because processes are not yet fully optimized within the digital infrastructure. To fully transition administrative workloads away from human staff and into Workday’s ecosystem, continuous process optimization has become a foundational priority for the company.
Envisioning the Future of Work: Proactive HR and Predictive Management
Once internal process optimization is fully achieved, Dow anticipates unlocking transformative benefits for both general employees and frontline managers. Rather than relying on traditional reactive notifications—such as a standard inbox alert indicating an upcoming performance review—future agentic systems could proactively anticipate user needs. For instance, a smart assistant integrated into Workday could recognize that an employee is logging in during performance review season, automatically aggregate check-in data tracked throughout the year, draft goal-alignment summaries, and interactively guide the staff member through the quantification of their annual contributions.
At the managerial level, embedded artificial intelligence promises to revolutionize workforce analytics. Instead of requiring leaders to manually run complex multi-variable attrition reports and cross-reference data across various business units, an agentic system could proactively flag subtle anomalies—such as an unexpected uptick in turnover among tenured personnel in a specific geographic segment—and prompt the manager to investigate before the issue escalates.
By shifting enterprise software from a passive repository of record into an active, proactive consulting tool, companies like Dow hope to eliminate unpopular administrative burdens, drastically reduce the time employees spend on routine paperwork, and generate reliable, quantifiable business insights. However, realizing this futuristic vision will ultimately depend on how successfully the software industry and enterprise customers can collaborate to resolve the foundational economic, governance, and regulatory challenges introduced by consumption-based agentic pricing models.
