Discover how Australian CTOs can set up AI for success, by scaling background AI, securing native APIs and governing digital assets.
Sydney Scott
Editorial Strategist, AI
Workday
On the floor at Workday Rising 2026 in Las Vegas, leaders were focused on the plumbing.
The real payoff in enterprise AI isn't sitting in a browser chat box; it’s running behind the scenes through headless connections, automated workflows, and strict governance.
The shift is direct: software is moving from manual prompts to event-driven execution. Leaders are deploying autonomous agents that run continuously in the background to handle multi-system operations without waiting for a user to type a request.
Scaling this takes more than raw model output. Success comes down to locking agentic autonomy directly to the exact security rules, approval chains, and business logic that run the company.
Here’s how top leaders are making that transition happen.
Guide
Companies often rush to show progress to boards and shareholders by chasing raw speed while skipping basic rules. Teams pull corporate data into outside systems or let digital agents run wild on raw files all for the sake of speed. But, this strips away the security checks, approval steps, and business rules that keep a firm safe.
While doing a demonstration at Rising, Workday Chief Technology Officer Gabe Monroy showed exactly how this plays out when he asked an unconstrained model connected to raw data to provide executive pay data. It simply read out the salary.
“You can export the data. You can't export the rules that govern the data.”
—Gabe Monroy, CTO, Workday
Monroy put the problem plainly, "You can export the data. You can't export the rules that govern the data."
Fixing this comes down to building governance directly into the connection layer instead of copying data into external systems. And by doing this automatically, organisations enforce existing permissions, approval chains, and business logic, balancing fast automation with reliable protection.
Enterprise implementation teams at Rising emphasised that getting your data in order is the best way to set up AI for success.
Global biopharmaceutical leader Sanofi makes the case: 12 years of technical debt left the company fragmented with 23,000 custom workarounds, 50,000 empty cost centers, 10,000 blank job profiles, and 441 business desks handling basic employee updates.
“It is essentially unusable," Danielle Bushen, the company's global head of people and culture tech strategy, told event attendees. Fixing the foundation, not the AI layer, is what made daily workflows run smoothly again.
Organising your business info—instead of stuffing job levels, pay grades, and location rules into a single notes box—gives software the clean inputs it needs to run smoothly. That clarity stops team members from having to play "human API," relying on insider knowledge to clean up information by hand.
"Think about this not as a technology project, but as a process improvement and data governance project,” Bushen advised. “Everything you need to touch is going to use the technology, but it is the process design and the data that it generates that will eventually get deployed through the tech."
“Think about this not as a technology project, but as a process improvement and data governance project.”
—Danielle Bushen, Global Head of People & Culture Tech Strategy, Sanofi
Putting this prep work first ensures agents do their jobs accurately and keep daily processes running effortlessly.
Relying on a chatbot interface for day-to-day business processes is an operational dead end. Forcing employees to type out prompts creates fatigue, wastes time, and leads to inconsistent inputs.
Phill Pearson, AI Center of Excellence Leader at Kainos, pointed out that up to 95% of proposed AI use cases shouldn't use a chatbot at all. They belong in background automation.
Now, leaders are adopting a concept inspired by Sakichi Toyoda—jidoka, or automation with a human touch. Automated systems run continuously in the background and pause only when human judgment is required.
In practice, these agent autopilots work quietly behind the scenes—monitoring shift changes, calendar updates, or expense submissions. The AI evaluates the event against company policy, completes routine updates automatically, and alerts a manager only when a policy violation or edge case requires human approval.
"Ambient AI working in the background is the best for adoption because people don't even know they're using it," Pearson highlighted.
“Ambient AI working in the background is the best for adoption because people don't even know they're using it.”
—Phill Pearson, VP, Workday AI Center of Excellence, Kainos
As companies roll out specialised AI agents across different departments, executive teams are adopting proactive management strategies to keep operations smooth and budgets predictable.
Three practical controls help leaders guide an AI workforce effectively:
Keeping architectures modular provides long-term agility as well. Jason Scheckner, vice president of AI go-to-market strategy at Workday, advised executives to design flexible systems that make upgrading tools simple.
"I'm betting on an agent today, but I'm going to remove that agent in six months because there's a new agent available," Schechner pointed out. "Does your architecture allow you to swap that agent out for the next one?"
“Does your architecture allow you to swap that agent out for the next one?”
—Jason Scheckner VP, AI GTM Strategy, Workday
For leaders turning these insights into an action plan, execution can be broken down in a 90-day plan:
Rising 2026 proved that while AI is evolving fast, the real work comes down to system design. To succeed, organisations will need to clean up their operational data, enforce strict security guardrails, and maintain total control over their digital workforce.
Guide
A gap between effort and impact is the real ROI problem. Discover the solution.