# See how a Salesforce org is really built, not how it was meant to be. > SyncOnAI 360 Architecture Intelligence derives how each Salesforce org is built from its own metadata: its automation mix, naming conventions, sharing posture and packaging. A canvas draws the data model, and a compact org profile is given to the AI so its answers and changes fit the existing design rather than a generic one. Source: https://synconai360.com/features/architecture ## Key facts - **4**: Dimensions profiled: automation, naming, sharing, packaging - **1**: Data model canvas per org - **Every**: AI request grounded in the org's profile - **12**: Audit sections, scored on their own ## Why Salesforce architecture is hard to see Architecture lives in diagrams and in people's heads. The org itself holds the truth, but reading it means opening hundreds of components and inferring the patterns by hand. - **Designs drift.** The documented architecture describes the org as it was planned, not as it is after years of change. - **Patterns are implicit.** Naming conventions and the preferred automation type are never written down; new work ignores them. - **The data model is unseen.** Object relationships live in Setup, one object at a time. - **AI ignores the design.** Generic AI recommends patterns your org does not use, adding another way of doing everything. ## What does a Salesforce architecture profile show? The Architecture lens describes each org as built: its mix of Flows, Apex, Workflow and Process Builder, its naming conventions, its sharing posture and how much of it comes from installed packages. It is derived from the org's own metadata on every sync, so it never goes stale the way a document does. The profile describes current state. It is used to keep recommendations consistent with existing practice, and it does not score or grade the architecture; scoring is the job of the health score and the audit. - Automation mix - Naming conventions - Sharing posture - Packaging footprint ## How do you see a Salesforce org's data model? The Canvas lens draws the object-level schema of each org, with a grid view and a dependency view, so the data model can be explored visually instead of object by object in Setup. The canvas focuses on the data model. For automation and code relationships, open a component in Lineage; for security, the audit's permission checks and the profile's sharing posture cover it. - Object-level schema - Grid and dependency views - Lineage for automation and code ## How does the AI respect an org's architecture? Every AI request carries a compact profile of the connected org: its edition, objects, conventions and scale. Answers and drafts are specific to that org rather than generic, and the chat checks what already exists before it builds. Org rules go further: they set naming conventions, forbidden patterns and the preferred automation type for each org, and changes the AI drafts are built inside them. The profile is a summary, bounded by what fits in a request; for detail the AI reads the repository directly. - Org profile in every AI request - Org rules the AI must follow - Existing solutions checked first ## How do you measure the quality of a Salesforce architecture? The audit scores twelve sections separately, including automation and Flows, permissions and security, Apex and code, limits and performance, packages and AI readiness, with evidence behind each finding. Findings like several Flows on one trigger, multiple triggers per object, profile-heavy security or sprawling fields point straight at architectural drift. Combined with the profile and the dependency graph, that gives an architecture review a current, evidenced starting point that can be shared as a read-only report. - Twelve scored sections - Findings that expose drift - Shareable read-only reports ## An architecture assessment 1. **Sync.** Connect the org and read its metadata. 2. **Profile.** See the automation mix, naming, sharing and packaging. 3. **Draw.** Explore the data model on the canvas. 4. **Score.** Read the audit's twelve sections and their findings. 5. **Govern.** Set org rules so new work, human or AI, keeps to the design. ## Architecture, documented versus derived | Without | With SyncOnAI 360 | |---|---| | Diagrams from the project kickoff | A profile derived from the org on every sync | | Conventions in people's heads | Conventions detected and enforced as org rules | | The data model one object at a time | A canvas of the whole schema | | AI suggesting foreign patterns | AI grounded in the org's own profile | | Reviews from opinion | Reviews from twelve scored sections | ## Who uses architecture intelligence - **Salesforce architects.** Start a review from the org as built: the automation mix, conventions, sharing posture and packaging, the data model on a canvas, and twelve scored audit sections, all current as of the last sync. - **Consultancies.** Assess an inherited client org in the first week and present an evidenced view of how it is really built, shareable as a read-only report the client opens without an account. - **Salesforce admins.** Set org rules for naming, forbidden patterns and the preferred automation type once, so every change the AI drafts follows the existing design. - **Platform teams.** Keep a long-lived org consistent as people come and go: the profile is derived on every sync, so it records how the org is built today rather than how a kickoff diagram said it would be. ## Frequently asked questions ### What is Salesforce architecture analysis? Understanding how an org is built: its automation, code, data model, security and conventions. SyncOnAI 360 derives this from the org's own metadata. ### Does the architecture profile score the org? No. It describes current state. Scoring comes from the health score and the audit's twelve sections. ### Can we see the data model? Yes. The Canvas lens draws the object-level schema with grid and dependency views. ### Does the canvas show automation? No, it focuses on the data model. Open a component in Lineage for automation and code relationships. ### How does the AI use the profile? A compact profile of the org is included with AI requests so answers and drafts fit the org's existing design. ### Can we enforce our conventions? Yes. Org rules set naming conventions, forbidden patterns and the preferred automation type, and AI-drafted changes are built inside them. ### How current is the profile? It is derived on every sync, so it reflects the org as of the latest refresh. ### Is architecture analysis on every plan? Yes. Every plan includes every feature. ### Can we share an architecture review with a client? Yes. Health and audit reports can be shared as read-only links that expire and can be withdrawn at any time, so a review can go to a client or steering group without giving them an account.