In a new episode of The Innovation Exchange, Workday's Yasmeen Ahmad talks with Accenture's Bharath Srinivas about what it takes for enterprises to move AI from insight to action—and why autonomy for AI agents has to be earned, not given.
For years, enterprise software has been measured by how well it captures the truth of a business—the system of record. But as AI agents move from answering questions to taking action on an organization's behalf, a new measure is emerging: can that same data and context be trusted enough to let software act, not just report?
That question sits at the center of a new episode of The Innovation Exchange, a Workday podcast exploring the people, partnerships, and real-world stories shaping Workday's ecosystem. In this conversation, host Yasmeen Ahmad, Workday's Senior Vice President and General Manager of Data Cloud, talks with Bharath Srinivas, Global CTO of Accenture's Workday Business Group, about what it actually takes for enterprises to move from AI pilots to AI that works autonomously in the flow of business.
Watch and listen to this full episode on YouTube, Spotify, or Apple Podcasts.
Srinivas has spent 14 years working with the Workday platform, but his fascination with technology started decades earlier, watching a computer from across a classroom in India.
That long view, he said, shapes how he thinks about today's pace of AI change. Clients aren't just adapting to new technology, they're confronting what he called an "existential" question about how to keep up.—His job is to help them demystify it and apply it to real work.
Srinivas's central theme: Workday customers are shifting from a system of record—the trusted source of what happened—to a system of action, where AI agents actually do the work. That shift, he said, depends on three things holding together at once.
"While a system of record serves as the definitive source of truth, an effective system of action requires three core pillars: unwavering trust in the underlying data, authorized agents empowered to execute tasks securely within defined limits, and, crucially, end-to-end explainability that connects what an agent did directly to why it did it."
Bharath Srinivas
Global CTO
Accenture Workday Business Group
In practice, that means bringing standardization to the workarounds and informal processes that often sit outside the system of record, so agents have something solid to act on. Without that foundation, Srinivas noted, "AI agents aren't like... super magical enough yet to understand all of that."
As Workday customers move from copilots—chat-based assistants a person directs—toward "autopilot" agents that act with more independence, Srinivas said trust has to be built in stages, not declared on day one.
It helps to separate two ideas here. Generative AI creates new content, such as text, summaries, or code, in response to a prompt, which is how most copilots work today. Ambient AI operates in the background, sensing context and acting on it without anyone asking, whether that’s drafting notes from a conversation or moving a candidate to the next step of a hiring process. The two often work together, with generative AI doing the creating inside an ambient system, but the shift Srinivas describes is largely a move from the first mode to the second.
He pointed to talent acquisition as the clearest example of this in action today. Accenture uses frictionless recruiting internally, automating a candidate's journey end-to-end based on policies already codified in Workday, with people stepping back in only for final reviews and decisions.
"We're now able to go in from step one to step ten, from our recruiting perspective, without human being in the loop. But the human is in the loop for the final reviews and making the decisions."
Bharath Srinivas
Global CTO
Accenture Workday Business Group
Where institutional knowledge isn't yet fully captured in the system—in moments like payroll close or converting a candidate into an employee—Srinivas said customers deliberately keep a human checkpoint, not out of caution for its own sake, but because people still hold context the system doesn't have yet.
Much of the current enterprise AI conversation focuses on pilots. The harder, less visible work, Srinivas said, is building what he calls the "last mile"—the roughly 40% of a process that standard agents cannot finish on their own, and where partners like Accenture step in.
To close that gap, Workday has expanded what forward-deployed engineering teams can build directly inside the platform through investments in Data Cloud, Developer Agent, and roughly 400 agent-ready tools. Rather than creating isolated customizations outside the system, partners can now orchestrate complex workflows natively.
Srinivas pointed to onboarding to show where partner impact really counts—bridging the gap between what an off-the-shelf agent can trigger and what a complete enterprise process actually requires:
"You think about that onboarding experience for an individual, which starts all the way from, 'Here's your offer... welcome to the organization,' to 'Here's your laptop,' walking into the building, and 'Here's your access.' Think of each of these steps as opportunities—handled for the most part by humans today, but tomorrow shared between AI agents, humans, or both."
Bharath Srinivas
Global CTO
Accenture Workday Business Group
Looking ahead, Srinivas connected Accenture's work with customers to Workday's recent focus on task-level visibility—understanding not just jobs and roles, but the specific tasks within them that AI can take on, freeing people to grow into new ones.
Asked to look further out, he raised the idea that the number of trusted, deployed agents could one day sit alongside headcount as a standard enterprise metric—dashboards where leaders dial up or down the mix of people and agents executing their strategy.
Explore more partner-built apps, agents, integrations, and services on Workday Marketplace—designed to extend the Workday platform and help organizations move, with confidence, from systems of record to systems of action.
More Reading
Workday is opening its platform to every builder. Build your first enterprise agent in minutes with Developer Agent, get your own development tenant (beta), and build your career and business in a thriving global ecosystem.
At Workday Rising, we announced that Agent Passport is in early access. Instead of asking whether an agent is safe, IT teams can now check a signed record of what it was tested for, by whom, and to what standard. Watch a quick demo.
Workday Extend AI Is Now Agent-Ready: Build With Speed, Keep The Trust
Developer Agent, Custom Agents, and Agent-Ready Tools are now generally available as part of Workday Extend AI—a comprehensive developer toolkit to build and run agents lawfully inside Workday’s guardrails.
