We recently spoke with Varun Gupta, VP of Engineering at Salesforce, and Sahil Vazirani, Principal Architect, to discuss how Temporal has become a foundational technology for Salesforce. Through Temporal’s orchestration capabilities, the company has significantly reduced manual intervention, optimized developer workflows, and accelerated its ability to deliver enterprise-grade solutions.
The Journey to Temporal: From Evaluation to Implementation
“Temporal is that glue that helps us stitch these complex systems together and support their interoperation.”
As one of the largest enterprise software providers, Salesforce operates a vast network of applications and infrastructure, much of it inherited through acquisitions. Managing this complex ecosystem posed significant challenges, particularly in coordinating distributed systems that are prone to failure.
“As we work to unite these systems, we need a common technology that helps them work seamlessly together to provide a unified experience for our customers,” Vazirani explained. “Temporal is that glue that helps us stitch these complex systems together and support their interoperation.”
The search for a solution began with thorough research. “I started reading more about the founders, Samar and Maxim, and realized the innovation they’d done in the industry, going back to 2009 with simple workflows, their work on Cadence, and how Cadence was open-sourced,” Vazirani recalled. His investigation included studying Cadence implementations at Uber and examining how Temporal evolved to address the limitations of previous orchestration systems.
Before Temporal, Salesforce relied on disconnected automation scripts and manual processes. Engineers had to oversee workflows in stages, executing individual steps across different systems. “For a large setup like a brand new region setup, it could take anywhere from days to months,” according to Gupta. The process required constant human intervention, with engineers needing to “press a button and then execute the next automation, execute the next automation.”
Building a Platform for Enterprise-Scale Automation
“What previously required 1,000 lines of code can now be done with just 50, thanks to the primitives provided by Temporal’s SDKs.”
The implementation started small but quickly demonstrated value. “One of our biggest challenges was finding a solution for running durable, reliable long-running workflows,” Vazirani explained. “Our existing solutions weren’t designed to support workflows of this duration.”
The impact on development efficiency was immediate. “What previously required 1,000 lines of code can now be done with just 50, thanks to the primitives provided by Temporal’s SDKs,” Vazirani shared. The ability to unit test code rather than work with domain-specific languages particularly improved the software development lifecycle and developer productivity.
As success stories emerged, information spread through Salesforce’s internal channels. “Our chief architect suggested to me, ‘Hey, this looks interesting and your recent success with the smaller use case allows us to unlock other possibilities,’” Vazirani recalled. He presented at the cross-cloud architecture forum, where other teams recognized similar challenges in their own work.
This momentum led to a strategic evolution in their approach. “We need to make this a real platform,” Vazirani explained. “How do we make it even further easy for our developers to avoid having the learning curve to learn a new technology, and start investing in developer productivity?” Their team began building an ecosystem of tools around Temporal, focusing on self-service capabilities, auto-debugging, and other features to support broader adoption.
Scaling for Enterprise Needs
“By reducing manual involvement and increasing automation, we’re making better use of our most valuable resource — our engineers.”
The transition to Temporal has transformed how Salesforce approaches workflow automation, but the team maintains a careful balance between expansion and reliability. “We want to enable every team at Salesforce to utilize this technology,” Gupta noted, “but we want to make sure that we are able to scale Temporal, and that we can handle the transaction volume at Salesforce’s immense scale, and serve our enterprise customers with the kind of performance they have come to expect.”
The impact on engineering productivity has been substantial. “We’ve seen tremendous productivity gains in engineering,” Gupta emphasized. “By reducing manual involvement and increasing automation, we’re making better use of our most valuable resource — our engineers. Instead of managing manual tasks and bespoke systems, they can focus on building core capabilities, developing features, and improving our infrastructure.”
The reliability improvements have been equally significant. Engineers no longer receive middle-of-the-night calls because they can trust their code will work. Even in failure scenarios, Temporal’s retry capabilities ensure processes resume from where they stopped, maintaining application reliability.
Future Vision
The scope of Temporal’s application continues to expand. “We are using it for a lot of infrastructure automation use cases, but now we are also expanding it to allow our applications to manage the application business workflows,” Gupta explained. This includes enabling inter-application workflows, where different applications need to interact to complete business processes.
For teams considering Temporal, Vazirani emphasizes the importance of starting small: “Don’t be intimidated by the technology. Try out the examples from the GitHub repositories — once you do, you won’t want to go back.” This approach of gradual adoption has proven effective for Salesforce, allowing them to transform their workflow automation while maintaining reliability at enterprise scale.
The team sees themselves as part of a larger community solving complex distributed systems challenges. “We are very invested in continuing to see this technology flourish,” Gupta noted. “We want to work with the community and increase community participation. We love being part of this collaborative effort to solve hard engineering problems.”
*This interview was conducted at Replay 2024, Temporal’s annual conference, where Gupta and Vazirani joined other industry leaders to share their experiences with durable execution and workflow orchestration.
Salesforce’s journey illustrates the broader adoption of Temporal across industries, as organizations seek reliable solutions for complex distributed systems challenges. Join the next generation of industry leaders at Replay 2025, taking place in London from March 3–5, 2025. Buy your ticket now to be part of the conversation and try Temporal for free today.*