Free AI tool · 30-day trial · No credit card
Which Salesforce integration pattern should you use?
Salesforce documents a handful of canonical integration patterns — Request and Reply, Fire and Forget, Batch Data Synchronization, Remote Call-In, UI Update, and Data Virtualization — and picking the wrong one is one of the most expensive mistakes in Salesforce architecture. A real-time callout where a nightly batch would do burns API limits and adds failure modes; a batch sync where users expect live data erodes trust in the system.
The right choice depends on a few concrete variables: data volume, latency tolerance, direction of flow, whether the source of truth lives in Salesforce or the external system, and how failures must be handled. Middleware questions layer on top — point-to-point HTTP callouts are fine for one or two integrations, but once you're connecting several systems, an integration layer like MuleSoft usually pays for itself in error handling, monitoring, and reuse.
This free tool asks for exactly those variables — source system, target system, volume, and latency requirement — and recommends a pattern with the rationale, credible alternatives, a MuleSoft yes/no call, and the watch-outs that typically bite in production. It's built on the patterns published at architect.salesforce.com, applied by AI to your specific scenario.
Integration Pattern Picker
Find the right architecture to connect Salesforce with any external system.
🔒 Describe scenarios in general terms — please don't include confidential client names or personal data.
How it works
Describe the two systems you're connecting, the expected record volume, and how fresh the data needs to be. The tool weighs those against Salesforce's documented integration patterns and governor limits, then returns a recommended pattern, the trade-offs versus alternatives, and specific risks to design around — like callout timeouts, API request limits, and duplicate handling.
When to use each pattern
- →Request and Reply — a user or process needs an immediate answer from the external system (e.g. real-time credit check).
- →Fire and Forget / Platform Events — Salesforce publishes a change and doesn't need to wait (e.g. order placed, notify ERP).
- →Batch Data Synchronization — large volumes on a schedule where near-real-time isn't required (e.g. nightly product catalog sync).
- →Remote Call-In — the external system creates or updates Salesforce records via API (e.g. a web form writing Leads).
- →Data Virtualization / Salesforce Connect — display external data in Salesforce without storing it (e.g. viewing ERP invoices on the Account page).
Do you need MuleSoft?
Not always. One or two simple integrations are often fine with native callouts and Platform Events. Middleware earns its cost when you have multiple systems, need centralized error handling and retry, transform data between formats, or want to reuse the same API across consumers. The tool gives you an honest yes/no with reasoning, not a reflexive middleware pitch.
Frequently asked questions
What are the main Salesforce integration patterns?
Salesforce's architect documentation defines six canonical patterns: Request and Reply, Fire and Forget, Batch Data Synchronization, Remote Call-In, UI Update Based on Data Changes, and Data Virtualization. Each fits a different combination of latency, volume, and direction of data flow.
How do I choose between real-time and batch integration in Salesforce?
Ask whether a user or process is waiting on the result. If yes, you need a real-time pattern like Request and Reply. If the data just needs to be consistent by some point in time, batch synchronization is cheaper, more resilient, and easier on API limits.
Is this tool free to use?
Yes — you get a 30-day free trial with no credit card required. Enter your scenario and get a pattern recommendation in under a minute.
Related tools
Try it now — free for 30 days
No credit card required. Part of Reliaarch's suite of 15 AI tools for Salesforce professionals.
Use the Integration Pattern Picker →