CRM & sales

Salesforce

Enterprise CRM with APIs, sandboxes, CLI, metadata, and automation.

Agents can automate CRM and metadata workflows, but enterprise permissions and object models require careful setup.

Agent access evidence

CLI

native

Salesforce CLI supports org auth, metadata deploys, data operations, and scratch orgs.

Evidence

API

native

Salesforce has REST, SOAP, Bulk, Metadata, and Tooling APIs.

Evidence

Browser

native

Setup UI is central for permissions, apps, and schema review.

Evidence

Account creation

strong

Developer orgs are self-serve; enterprise orgs are contracted.

Evidence

Pricing clarity

strong

Public editions and add-ons; enterprise pricing and API limits can be complex.

Evidence

Docs quality

strong

Developer docs are broad but require platform expertise.

Evidence

Sandbox

native

Sandboxes and scratch orgs support safer agent testing.

Evidence

Readiness score

Strong

Strong: 83/100. Strongest in agent safety; weakest in production reliability.

Machine operability21/25

Agent safety23.2/25

Agent readability17.4/20

Auth and setup12/15

Production reliability9.4/15

Best for

Enterprise CRM, metadata deployments, customer records, integrations, and sales operations.

Pricing

Public editions and add-ons; enterprise pricing and API limits can be complex.

Auth

OAuth, connected apps, JWT bearer flow, profiles, permission sets, and org scopes.

Account creation

Developer orgs are self-serve; enterprise orgs are contracted.

Browser support

Setup UI is central for permissions, apps, and schema review.

CLI

API

https://login.salesforce.com/services/data

Limitations

Key limitations: Production org writes, PII, workflows, and metadata deployments need admin review. Verify official docs, pricing, rate limits, permissions, and production side effects before delegating account-changing actions to an agent. CLI access is powerful but can run destructive local or production commands; prefer dry runs, preview environments, and non-production credentials. API automation needs client-side throttling, idempotency where supported, pagination limits, retries, and audit logging.