Explore code and architecture intelligence

Overview

Org. Code & Architecture Intelligence starts with the org-wide picture: how the org is built, which patterns hold, and which ones get bypassed. Below that, you can analyze any single component in depth. Open it from the sidebar under Org Inventory → Org. Code Intelligence.

Every AI analysis is saved and shared with your workspace. A component that someone has analyzed once stays analyzed for everyone.

Before you start

• The org must be synced and indexed. This page reads the local index. See Index an org.
• Any signed-in workspace member can read the page. Rebuilding the index needs the Index / re-index orgs permission, which every built-in role except Viewer has.
• The Architecture story and Analyse this component need an AI provider. All measured figures are free and work without one.

Steps

1. Pick an org from Select an org…. If its index looks stale, click Index.
2. Read the signal tiles: Components, Dependencies, Functional areas, Entry points, External systems and Cycles. Hover over a tile to see how it is measured.
3. Click Generate briefing under Architecture story for a written narrative of the architecture. It is saved, and you can Regenerate it later.
4. Work through the panels. Use Expand all or Collapse all to open or close them together.
   • Major functional areas: clusters of components that work together.
   • Highly connected components: components ranked by significance. This is not a defect score.
   • Critical execution paths: the main route out of each entry point.
   • Patterns & anti-patterns: each pattern is graded Anti-pattern, Inconsistent, Not present or Healthy.
   • Apex test coverage: click Check coverage to read coverage live from Salesforce. It uses API calls, so it never loads on its own.
   • Where the data gravity is and Integration touchpoints.
   • What deserves attention: each item is tagged Observed, Inferred or Worth checking, depending on how firmly the metadata supports it.
5. In the component browser, search with Search components, objects, domains…. Filter by type, or switch between All and Analysed. Results are ranked by significance, and the list shows up to 400 components.
6. Click a component to open its dossier: Significance, Called by, Calls, Blast radius, Betweenness, Lines, SOQL, Complexity and Governor risk.
7. Click Analyse this component for an AI analysis. Switch between the Executive, Functional, Technical and AI Assessment views.

Example

Priya is planning a refactor in acme-uat. She searches for AccountService and sees that it has a large Blast radius and high Betweenness. That tells her many execution paths depend on it. She clicks Analyse this component and reads the AI Assessment view. Then she adds the risks to ticket ACME-142 before any code changes.

FAQs

Is significance the same as risk?

No. Significance combines reuse, blast radius and brokerage. A highly significant component isn't necessarily bad. It's simply important, so change it with care.

Does Check coverage run my tests?

No. It reads the latest stored coverage results from the org. A label shows how recent those results are, such as Measured recently, Ageing or Stale.

Why does a saved analysis show a warning?

The component may have changed since the analysis was written. Click Re-analyse to refresh it.

Troubleshooting

"acme-dev has no analysable components indexed."

The org hasn't been synced and indexed yet. Click Index now. See Pages show no data.

"Could not read coverage from this org."

The org's stored credentials may have expired. Reconnect the org and try again. See Fix expired org credentials.

"Could not analyse that component. Try again."

The AI provider failed or isn't configured. Check AI provider not configured.

Related articles

• Index an org
• Understand an org with Org Overview
• Scan Apex, triggers and Flows with Code Analyzer
• Map automation and async execution