Salesforce Claude Code engineering agent
Salesforce engineering teams use Claude Code to reason over the company codebase, implement changes, and execute API migrations.
30 cases where an agent handles software engineering work, 24 of them with reported numbers. Companies include Nubank, ADP, Amplitude, Carvana.
Cases with reported numbers first, then newest.
Salesforce engineering teams use Claude Code to reason over the company codebase, implement changes, and execute API migrations.
Simplyhealth engineers and administrators use Claude to analyze Salesforce code, map dependencies, find duplicate logic, and explain errors.
Amplitude teams use Wave to identify product opportunities, propose changes, route plans to coding agents and measure outcomes.
CodeRabbit uses coding and security agents to produce software, review generated code, triage incidents, and prepare fixes for human approval.
CTBC Bank uses a governed multi-agent pipeline to recover business rules from COBOL and generate specifications, SQL, and tests.
TCS uses a production agent system to turn BFSI requirements into governed API and browser tests, execute them, and classify failures.
Circle K agents investigate infrastructure and application signals after alerts and send a structured incident summary to on-call engineers.
Eon's SRE agent investigates production across AWS, Azure, and Google Cloud through a centrally authorized gateway of operational tools.
Epsilon product teams describe browser tests in plain language, and a managed agent executes the steps and records evidence and results.
MoEngage exposes private regional APIs to agents as permission-scoped tools without public endpoints or model-controlled identity fields.
Nubank's agent gathers recovery evidence from metrics, logs, and alerts and produces audit-ready reports for disaster recovery tests.
Rescale's production agent classifies nightly end-to-end test failures, gathers supporting logs, and recommends fixes for engineer review.
Katalon agents analyze requirements, create and maintain tests, execute them, detect defects, and surface findings with human oversight.
Carvana uses an agent to investigate Slack alerts, identify code owners, draft root-cause reports, and open pull requests.
Coupa's Sentinel agents triage test failures, investigate incidents, open pull requests, and answer engineering support questions.
Sentry hands Seer's root-cause analysis to a coding agent that implements a fix and opens a pull request.
Delight.ai uses Claude Code to build debugging and regression tooling for customer-support agents before deployment.
A global bank uses Sapient Slingshot on Google Cloud to turn COBOL code into traceable specs and backlog items.
RUSH uses MuleSoft Vibes as an AI coding collaborator for healthcare integrations, DataWeave transformations, and API scaffolding.
Crossmint used Devin to reduce the integration backlog for GOAT SDK blockchain tooling.
Gumroad uses Devin across engineering, product, support, and marketing to turn requests into merged PRs.
Litera embedded Devin agents into engineering teams to automate quality engineering and regression work.
Nubank used Devin to parallelize migration of 100,000 data class implementations, achieving 12x efficiency and 20x cost savings over manual work.
Ramp's internal AI agent monitors Sentry for top production errors, writes fix PRs autonomously — now behind 30% of merged PRs.
ADP engineers use AWS Transform agents to extract tax business rules and dependencies from legacy systems for a configuration-driven replacement.
Intuit's EWOK Agent turns plain-language failover requests into policy-checked recovery workflows across infrastructure tiers.
PwC uses CrewAI agents to generate and validate proprietary-language code and long technical specifications.
Builds HTML Kanban with drag-drop, localStorage, dark mode in 60s.
Creates cards for issues, manages roadmap.
Builds functional prototypes for growth ideas with team iteration.