
By Aya Yohanan-Rokach and Vivienne Wei.
How do you enable thousands of engineers to adopt AI tools at a 94% rate while maintaining enterprise security and reliability? Salesforce achieved this by operationalizing AI productivity at scale and fostering an AI-native engineering culture, with a critical focus on operational trust. “Trust is our number one value at Salesforce….every transaction and every conversation at Salesforce begins and ends with the word trust,” said Srini Tallapragada, President & Chief Engineering Officer and Customer Success Officer at Salesforce.
The Unified Agentforce Platform (UAP) acts as Salesforce’s all-inclusive AI hub, providing fully customizable autonomous AI agents that seamlessly connect and act on enterprise data across sales, service, marketing, and commerce. UAP integrates Data Cloud for data storage and retrieval, AI Cloud for advanced LLM implementations, MuleSoft for connectivity, and robust analytics capabilities, creating a cohesive platform that supports all Salesforce clouds.
However, building this foundation involved more than just technical integration. It required a fundamental shift in how engineering teams approach operational excellence, security practices, and trust at an enterprise level.
From Trust Evolution to Systematic Change Management
In 2024, Salesforce engineering leadership identified an opportunity to strengthen operational foundations across the organization. An uptick in high-severity incidents highlighted areas for improvement, leading to a strategic “Trust Moratorium”— a focused period dedicated to enhancing operational excellence before resuming feature development.
This initiative tackled systemic challenges that arose as the organization scaled. Many manual changes were occurring monthly, presenting opportunities for process improvement. Different clouds had developed independent practices, and security initiatives often disrupted planned development work with urgent timelines.
UAP took a leadership role in implementing the organization-wide Safe Change Program, revolutionizing deployment processes across all engineering teams. The systematic approach addressed the root causes of operational instability through comprehensive automation requirements:
- 100% automation for deployment processes
- 80%+ code coverage maintenance
- Automated FIT testing
- 30-minute pipeline execution times
- 100% automation for certificate and secret rotation processes

Chart illustrating adoption of new enhanced quality controls.
This initiative wasn’t just about enforcing rules; it was about building infrastructure that eliminated human error and enabled rapid, reliable, automated deployment cycles. The results were immediate and measurable: manual changes dropped dramatically and 100% managed release adoption was achieved for Hyperforce services.
The automation infrastructure created a stable foundation, allowing engineers to experiment confidently with AI tools, knowing that deployment pipelines would catch issues before they reached customers.
Building Cross-Cloud Security Collaboration
In 2024, UAP’s security practices across multiple clouds had become fragmented, with MuleSoft, AI Cloud, Analytics, Platform and Data Cloud teams operating independently and duplicating efforts. This lack of unified standards hindered visibility into the security organization’s roadmap and priorities.
To address this, the UAP Security Council was formed as a cross-cloud virtual team of engineering and security stakeholders. The council’s goal was to foster a culture where service owners actively “own their Trust.” This framework not only established security best practices but also promoted collaboration, knowledge sharing, and eliminated redundant efforts.
The results were significant:
- A 62.85% reduction in bugs initially classified as P0 to P1 through risk-based scoring algorithms.
- A 52% reduction in P0-P4 vulnerabilities in non-Hyperforce environments during Q1 FY26.
By providing clear visibility into security roadmaps and creating forums for sharing successful mitigation strategies, the council enhanced development velocity.
Scaling Trust Practices Across Thousands of Engineers
The most significant challenge UAP faced wasn’t technical but cultural. As the organization doubled in size and expanded across continents, scaling trust practices across diverse engineering backgrounds and experience levels became increasingly complex.
To address this, UAP implemented a shift-left ownership culture, empowering service owners to integrate trust practices directly into development workflows. This transformation included:
- Establishing security guide programs to upskill engineers.
- Creating clear deliverables that define required guardrails before product release.
- Building collaborative cross-cloud engagement models.
Specialized councils—such as the Security Council, Availability Council, and Cost-to-Serve Council—formed virtual cross-cloud teams to share best practices and drive excellence in their respective domains. This systematic approach ensures service owner accountability and provides the necessary tools and knowledge for delivering trustworthy products from the start.
The cultural shift has enabled rapid AI experimentation, leading to a 94% adoption rate. When engineers trust their security guardrails and deployment automation, they can iterate quickly with AI tools without fear of introducing production instabilities.
The Visibility Revolution: Engineering 360 Dashboard
Operational excellence demands clear visibility, but UAP’s distributed architecture across multiple clouds led to scattered monitoring, hindering comprehensive health tracking. Teams lacked a clear understanding of their trust maturity, risk exposure, and progress toward operational excellence targets.
The Eng360 Dashboard powered by Salesforce Data Cloud and Tableau addressed this challenge by providing a centralized, single source of truth for security, availability, quality, accessibility, and developer productivity metrics across all UAP services. This comprehensive monitoring platform enables service teams to track Safe Change compliance through SCC scorecards, maintaining real-time deployment readiness status.
Regular engineering health reviews use Eng360 data for weekly, monthly, and quarterly assessments with senior leadership, enabling proactive risk management and continuous operational improvements. This comprehensive visibility allows engineering teams to confidently use AI tools while maintaining enterprise-grade reliability, proving that operational excellence is not a constraint on innovation but a foundation for sustainable growth.
Learn more
- Stay connected — join our Talent Community!
- Check out our Technology and Product teams to learn how you can get involved.