How Zuora unified AI across 200+ SaaS apps with orchestration

Zuora, an industry leader in solutions for modern recurring revenue businesses, runs more than 200 SaaS applications across its enterprise landscape. As AI capabilities surfaced across those systems, each introducing its own interface, assistant, and logic, Zuora recognized an emerging enterprise risk: AI fragmentation at scale.

“If an employee has to go to each of the SaaS applications and interact with their AI agent, it is going to be a fragmented and disjointed experience,” said Zuora CIO Karthik Chakkarapani.

Zuora did not want isolated AI tools scattered across the stack. It wanted a unified experience.

“It’s almost like someone walking into a hotel, and they see multiple concierge desks,” Karthik continued. “Imagine that experience versus going to a single concierge desk in the hotel and getting all your needs addressed.”

At the same time, Zuora’s integration layer was holding the business back. Integrations were distributed across MuleSoft, custom AWS code, and other tools. Delivery cycles stretched into months, limiting the speed and flexibility a modern SaaS business requires.

Mark Gill, Head of Infrastructure and Services at Zuora, explained, “People didn't think about our integration layer as a place to solve problems. It was too hard.”

If Zuora was going to scale AI responsibly, it needed orchestration, not more fragmentation.

Zuora standardized on Tray as its enterprise orchestration platform for data and AI. The platform brings together agent development, intelligent automation, AI-ready integration, and unified governance across systems, creating a single, governed layer behind every AI-driven experience.

The goal was not simply faster integrations. It was control, visibility, and adaptability across systems. Instead of long development cycles and monolithic builds, the team adopted an iterative model. Ideas were prototyped quickly, workflows were validated in production, and iterations followed.

That shift changed how the business viewed the integration team. What had once been seen as a bottleneck became a partner to sales, marketing, and operations. Instead of being asked why something couldn’t be done, the team was now delivering new solutions in days and collaborating directly with business units on experimentation and execution.

“We can have integrations up and running faster than ever before,” said Mark. That shift in speed changed how the business viewed IT.

What began as an integration modernization effort became the foundation for enterprise AI orchestration, letting the team move beyond integration projects and deploy AI-powered workflows in production.

One of Zuora’s most impactful AI workflows focuses on customer briefings.

Before meetings, sales and customer success teams needed a consolidated 360 view of account activity across CRM, support, usage, and other systems. Preparing a single executive brief required approximately 23 hours of combined manual effort.

Using Tray, Zuora built an AI-driven workflow that orchestrates this process end-to-end. A user initiates a request in Slack. Tray identifies the relevant systems, retrieves structured data, applies AI reasoning to synthesize insights, and generates a formatted executive brief.

What previously took 23 hours now takes under five minutes.

The workflow processes multiple briefs per week and reclaims 170+ hours weekly, equivalent to 2–4 full-time employees’ capacity.

The team now spends time interpreting and acting on insights rather than gathering data.

The customer briefing workflow was not an isolated automation. It demonstrated how a centralized orchestration layer could connect systems, apply AI reasoning, and enforce governance in a controlled way.

Instead of embedding AI separately inside each SaaS application, Zuora routed workflows through a unified orchestration layer. Permissions remained consistent with existing access controls. Executions were observable and integrations were reusable.

That model now serves as the foundation for additional AI-driven workflows across the business.

Zuora’s vision goes beyond individual use cases. The company is building what it describes as an “agent of agents” model: a single AI interface that routes requests intelligently across systems.

“We want a simple framework orchestrated so that all the magic happens behind the scenes,” says Karthik.

Instead of turning on AI inside each application and hoping for consistency, Zuora centralized orchestration. Systems connect through a governed integration layer. AI reasoning happens within defined boundaries. Workflows are observable, auditable, and extensible.

The result is measurable operational impact today and a scalable foundation for multi-agent orchestration tomorrow.

Watch the on-demand webinar with Zuora CIO Karthik Chakkarapani to learn how they built scalable, production-ready AI agents with governance and control.