Salesforce Platform guide
A practical guide to the Salesforce Platform, the shared foundation for custom objects, automation, code and apps across every Salesforce cloud.
The Salesforce Platform is the shared foundation under Sales Cloud, Service Cloud and most other Salesforce products. Your data model, security, automation and custom code all live there. Salesforce's product page calls it Salesforce Platform. Blogs and job adverts still use older names such as Lightning Platform and Force.com, and some Salesforce material from 2025–26 says "Agentforce 360 Platform" or "Headless 360 Platform".
What it is for
Use the platform when standard CRM features get you most of the way but your processes, data or user experience need more. Typical work includes:
- custom objects for a business process
- record-triggered automation and approval routing
- integrations with back-office systems
- internal apps that share the same users, sharing rules and audit history as the rest of your org
Core capabilities
| Building block | What it does | Current notes |
|---|---|---|
| Lightning App Builder | Drag-and-drop record, app and home pages | Declarative, so admins can own it |
| Flow | Flow Builder, Flow Approvals and Orchestration | Winter '27 adds Test Mode (Beta) and a User Context run mode |
| Apex | Server-side code for complex logic, integrations and bulk jobs | Winter '27 raises heap limits to 10 MB sync and 25 MB async |
| Lightning Web Components | Standards-based UI components | Complex template expressions and lwc:external are GA in Winter '27 |
| Salesforce Multi-Framework | React apps on the platform | Angular support added in Winter '27 |
| Headless Experience Layer | Define an experience once and render it in Agentforce, Slackbot and other surfaces | Rolling out; Salesforce sources differ on Beta or GA |
| DevOps Center | Change tracking and deployment pipelines | The Winter '27 platform blog calls it Lifecycle Center |
| Agentforce Vibes | AI-assisted coding and app building | Review generated code like any other |
How it connects
Every core cloud runs on the platform. A custom object or Flow you build is available to sales, service and portal users alike. Agentforce agents can use platform automation and code as actions, and Data 360 adds unified data for automation and grounding. Salesforce's Headless 360 initiative opens the same platform to AI agents through MCP servers, APIs and the CLI. When a workload does not fit inside the org, it usually moves to Heroku or goes through MuleSoft. Security add-ons sit in Shield and Salesforce Guardian.
Who it suits
It suits organisations whose processes don't match standard CRM, and those with an admin or developer to look after what gets built. Small teams on standard processes often need only configuration.
Planning an implementation
- 01
Map processes and data
List the objects, owners and volumes before you design screens or automation.
- 02
Choose declarative or code
Start with Flow. Use Apex when logic, volume or testing needs justify it.
- 03
Design the data model and sharing
Decide record ownership, sharing and field-level security early.
- 04
Set up source control
Use DevOps Center or the Salesforce CLI with a sandbox pipeline, not changes made straight in production.
- 05
Plan for releases
Three releases a year change behaviour. Winter '27 is API version 68.0.
- 06
Review integration authentication
Salesforce says connected apps support ends in Summer '27 and points to External Client Apps. Check Help for your org.
Common mistakes
- Stacking several record-triggered flows and triggers on one object with no agreed order.
- Hard-coding IDs, endpoints or credentials instead of using Named Credentials.
- Treating AI-generated code or flows as finished without review and tests.
How GeekMidst can help
We can help you decide what to configure and what to build, then deliver it. Our Salesforce development services cover Apex and integrations, our Flow automation services cover process design, and our LWC development covers custom UI. For larger programmes, architecture consulting helps set standards before the build starts.
Questions we hear
Is Lightning Platform still the product name?
Should we use Flow or Apex?
What is Headless 360?
Related reading
Salesforce Flow vs Apex
When to use declarative automation and when to write code
→Record-triggered Flow best practices
Patterns that keep automation fast and predictable
→LWC data patterns
Working with Lightning Data Service and Apex
→Heroku guide
When to run workloads outside the org
→Sources
Planning custom work on Salesforce?
Tell us what you need to build and we will help you choose between configuration, Flow and code.
