Automation at Scale
Conquering Salesforce login complexity to deliver a rock-solid test suite — from a fragile 6-hour run to a fast, trustworthy CI pipeline.
Automation engineering for an integration-heavy sales productivity platform.
The client is an AI-powered sales productivity platform that operates directly inside Gmail, Outlook, and Microsoft Teams. It bridges the gap between email inboxes and Salesforce CRM, eliminating manual data entry and empowering sales teams to book more meetings, track buyer signals, and maintain a cleaner pipeline — all without leaving their inbox. The platform serves over 50,000 customer-facing teams worldwide, including several Fortune 500 enterprises and recognisable consumer brands. Its deep integrations with Salesforce, Google Workspace, and Microsoft 365 make it one of the most integration-intensive SaaS products to test and automate.
Automation for deeply integrated SaaS products requires architectural thinking, not just script writing.
Automated Test Cases
Execution Time Reduced
FULL SUITE COMPLETION
Flaky Test Rate
When login itself is the hardest test.
Automating quality engineering here meant testing an application deeply entangled with Salesforce, Gmail, and Outlook — platforms with strict, automation-resistant security policies that made a conventional automation setup unworkable from the start.
Salesforce Auto-Lock
Salesforce locks accounts after repeated or failed login attempts. With multiple parallel test scripts authenticating simultaneously, account lock-outs were constant — halting entire test runs before a single test could execute.
Multi-Platform Authentication
The application requires concurrent authentication across Salesforce, Gmail, and Outlook browser extensions. Each platform uses different OAuth flows and session behaviours, making a unified auth strategy extremely complex.
6-Hour Execution Time
The initial full-suite run took over 6 hours due to sequential execution and repeated login cycles for every test — far too slow for a continuous delivery model with weekly production releases.
Widespread Test Flakiness
Login-related failures and inconsistent browser extension loading caused pervasive test flakiness. Tests failed intermittently without any code change, eroding confidence in the entire automation suite.
A session-level fix, not just more scripts.
The team designed a multi-layered automation strategy that addressed each challenge systematically — combining a shared session architecture, parallelised execution, AI-assisted code development, and robust OAuth handling across all three platforms. Midway through the engagement, the team's AI coding workflow migrated from Cursor AI to Claude, alongside a purpose-built Claude skill that connects directly into the client's TestRail instance.
Shared Session Storage
Authentication tokens and cookies are established once and shared across all test files — eliminating the root cause of lock-outs (repeated logins per test) without sacrificing test isolation.
Project-Level Cookie Architecture
Native login bugs were fixed and a project-level cookie management system was built, allowing authentication state to persist and be reused across runs — five framework fixes merged to the development branch.
Parallelised Test Execution
Project-level parallel execution was enabled for the entire suite. Specs that previously ran sequentially over 6 hours were restructured to run in parallel, cutting total execution time to under 2 hours.
2FA & OAuth Automation
Two-factor authentication was automated across all scripts to handle activation flows triggered by automated logins. Email OAuth flows, including the complete Forgot Password journey, were fully automated with lint compliance.
AI-Assisted Development with Claude
The team's AI coding workflow migrated from Cursor AI to Claude, now used throughout the automation lifecycle to accelerate code generation, suggest refactors, and fix lint compliance — reducing time-to-automation for each new test case and improving code quality.
Production Regression & Smoke Suites
Structured manual production regression was delivered for every weekly release, alongside a dedicated smoke test suite for rapid post-deployment validation.
Claude Skill for TestRail-Integrated Test Case Generation
Built a custom Claude skill that helps QE engineers author test cases directly against TestRail. A TestRail MCP integration fetches existing test cases for AI-assisted coverage analysis, then generates additional test cases through a read-only workflow that never modifies TestRail directly.
What the work delivered.
Salesforce account lock-outs eliminated
the core blocker that previously caused entire test runs to fail was resolved through shared session storage and project-level cookie management.
Full test suite execution reduced from over 6 hours to under 2 hours through parallelisation
a 66%+ improvement enabling rapid CI/CD feedback cycles aligned with weekly production releases.
350+
automated test cases now passing consistently with a near-zero flaky rate, providing a stable regression safety net for every release.
End-to-end multi-platform authentication fully automated
covering two-factor auth, email OAuth, and Forgot Password flows across all environments.
AI coding workflow migrated from Cursor AI to Claude, accelerating automation development velocity and reducing time-to-automation per test case through AI-assisted review and refactoring.
A custom Claude skill built for TestRail-integrated test case generation
using a TestRail MCP integration to fetch existing test cases for AI-assisted coverage analysis, then generating additional test cases through a read-only workflow that never modifies TestRail directly.
A custom QE dashboard delivered early in the engagement, providing real-time visibility into manual vs. automated coverage, pending automation backlog, and non-automatable scenarios.
The stack.
By solving the authentication problem at the session level and restructuring execution for parallelism, the team transformed an unstable, 6-hour test suite into a fast, reliable automation foundation that confidently supports every weekly production release.