Sandeep Kumar Panda
Designed

An AI engineering operating system for a Salesforce practice

A working model for how a Salesforce team uses AI across three surfaces: think, build and reach.

Role
Author

How it fits together

  1. Verify the org

    Identity confirmed before any work begins

  2. Think

    Discovery, analysis, architecture, documents

  3. Build

    Repository, metadata, tests, deployment

  4. Reach

    Orgs where no API access is granted

Orange marks the parts I built.

The model

A working model for how a Salesforce team uses AI across three surfaces: one to think (discovery, analysis, architecture, documents), one to build (repository, metadata, tests, deployment), and one to reach orgs where no API access is granted.

What it defines

  • Where each kind of knowledge lives, so nothing important sits only in a chat.
  • How context crosses between tools without being retyped.
  • The rule that org identity is verified before any work begins.

Skills used

  • AI operating model
  • Delivery process
  • Knowledge management

Next case study

A discovery and delivery pack for a marine aquarium retailer