SI and AI services · Service
Claude Code for Salesforce and HubSpot development
Set up Claude Code so your admins and developers ship Apex, Lightning Web Components, Flows and HubSpot customizations faster, through the same source control, tests and reviews you use today.
Short answer
Claude Code is Anthropic's agentic coding tool. In a source-tracked Salesforce DX or HubSpot project, it can read metadata, draft Apex, Lightning Web Components, tests and deployment scripts, and explain an unfamiliar org. We set it up with the right permissions, project instructions and guardrails, and fit it into your review and deployment pipeline.
This service sits in Step 3, workflow redesign of our SI and AI deployment sequence. See all six steps.
Most Salesforce teams have more change requests than developers. Claude Code helps by doing the first draft of the work: the Apex class and its test, the LWC, the Flow documentation, the deployment manifest. A person still reviews every change, and nothing reaches production except through your pipeline.
What makes this safe is the setup around it: sandbox-only credentials, project instructions that encode your standards, required tests and code review.
What we set up
| Area | Setup |
|---|---|
| Project | Salesforce DX project with source tracking, or HubSpot CLI project, in your repository |
| Access | Sandbox and scratch-org credentials only; no production deploy rights for the tool |
| Instructions | A project file with your naming, security and testing standards so output matches house style |
| Tools | Salesforce CLI and MCP servers your security team approves; read-only where possible |
| Quality | Required unit tests, static analysis, and pull-request review before merge |
| Admin | Claude Code under your Team or Enterprise plan, with usage visible to admins |
What you get
- Repository and project setup for Salesforce or HubSpot
- Project instructions file and reusable skills for your common tasks
- Guardrails: sandbox-only access, approved tools, required tests
- Pipeline fit with your CI and deployment process
- Hands-on sessions with your admins and developers on real backlog items
How it comes together
- Review your pipeline. Source control, environments, tests and approvals.
- Set up the project. Credentials, instructions and approved tools.
- Pair on real work. Three to five backlog items done together.
- Tighten. Update instructions from what the reviews caught.
- Hand over. Your team runs it; we support through a retainer if you want.
Is this the right call?
Good fit when
- Salesforce or HubSpot teams with a backlog and source control in place
- Admins who want help with Apex tests, Flow documentation or metadata clean-up
- Firms moving from change sets to a source-driven pipeline
Think twice when
- No source control or sandbox discipline yet; fix the pipeline first
- Work that is mostly declarative setup with little code; a senior admin may be the better spend
Mistakes we help you avoid
Giving the tool production credentials
Keep it in sandboxes. Production changes go through your pipeline and a person.
Skipping tests because the code looks right
Generated code still needs tests that prove behavior, not just coverage.
Questions we get
Can Claude Code work on Salesforce?
Yes. In a Salesforce DX project with the Salesforce CLI, Claude Code can read metadata and draft Apex, Lightning Web Components, tests and deployment files. Changes should go through code review, tests and your deployment pipeline.
Is it safe to point an AI coding tool at our org?
It is when it works in sandboxes with limited credentials and every change goes through review. We set it up that way and keep production deployment with your pipeline.
How is Claude Code related to Agentforce Vibes?
Agentforce Vibes is Salesforce's own development experience, and under Claudeforce Claude powers it by default. Claude Code is Anthropic's tool that works across any repository. Many teams use both.
What does this cost?
Setup and enablement are quoted from your team size and pipeline. Claude Code usage is part of your Anthropic plan or API billing.
Free, about 3 minutes: take the SI and AI readiness quiz. Answer 10 questions and get an instant readiness score across 10 areas. No email needed to see your results.
Updated October 7, 2026. Vendor features and terms change; we confirm them for your org before scoping.
