50 Platform Integration Architect Practice Questions: Question Bank 2025

A company is assessing its current integration landscape. Several upstream systems publish customer updates, but no one can reliably answer where the “system of record” is for Address and Phone. What is the BEST first step for an Integration Architect to reduce long-term integration risk?

A retail company needs near real-time updates from Salesforce to an external inventory system whenever Order status changes. The integration must be loosely coupled and allow multiple downstream subscribers in the future. Which approach BEST fits these requirements?

An external system must create and update 5 million records in Salesforce weekly. Data can be up to 24 hours old, and the primary goal is throughput with minimal API overhead. Which Salesforce API is the BEST fit?

A Salesforce org calls an external shipping service to get real-time rates during checkout. Occasionally, the service times out. The business requires that users can still complete checkout by selecting a default shipping method when the service is unavailable. What is the BEST integration design principle to apply?

A global company needs a single customer profile across Salesforce and multiple ERPs. Each ERP owns different attributes, and Salesforce owns preferences. The solution must resolve duplicates and publish a consolidated “golden record” to subscribers. Which pattern BEST meets this requirement?

An integration must upsert Accounts from an external system into Salesforce. The external system has a stable unique identifier and wants to avoid creating duplicates. What is the BEST approach in Salesforce?

A customer implements an event-driven integration where Salesforce publishes events and an integration middleware subscribes. Operations reports that some downstream systems occasionally miss updates during maintenance windows. Which capability BEST addresses this while keeping the architecture event-driven?

A Salesforce-to-middleware integration uses a Named Credential. Calls fail with 401 Unauthorized after the security team rotates the external system’s credentials. What should the Integration Architect recommend to reduce future outages from credential rotation?

A company must integrate Salesforce with an on-premise system that is not reachable from the public internet due to strict network policy. The integration must support near real-time request/response from Salesforce to on-prem services. Which solution is MOST appropriate?

A Salesforce org publishes events that trigger updates to an external billing system. The business requires that a billing update occurs only if the Salesforce transaction commits successfully, and they also want to avoid duplicate downstream processing when Salesforce retries or users update the record multiple times quickly. Which design BEST satisfies these requirements?

An integration team is assessing an existing landscape with Salesforce, an ERP, and several custom apps. They need to quickly determine which systems are systems of record for key business entities to avoid conflicting updates. What is the BEST first step?

A business stakeholder requests near-real-time updates in Salesforce when an order is shipped in the warehouse system. The warehouse system cannot accept inbound calls but can publish messages to a broker. Which integration requirement should the architect capture to guide the design?

A company needs an integration to create a Case in Salesforce from an external web portal. The portal requires an immediate success/failure response to show the user. The expected volume is low, but the portal must not expose Salesforce credentials. What is the recommended approach?

An enterprise wants to standardize integrations across Salesforce and multiple back-end systems. They require consistent security, throttling, schema validation, and observability. Which approach best meets these needs?

Salesforce must send customer updates to three downstream systems. Each system has different payload needs and occasionally becomes unavailable. The business requires that updates are not lost and that systems can catch up after downtime. Which design is BEST?

A partner integration uses Salesforce APIs and intermittently fails with HTTP 401 errors. Logs show the access token is valid at the start of the day but requests later fail. The partner claims they "reuse the same token" to reduce overhead. What is the MOST likely root cause and fix?

An architect is defining integration requirements for a bidirectional Account sync between Salesforce and an MDM. Users can update Accounts in Salesforce, but MDM is the system of record for legal name and tax identifiers. What requirement best prevents update conflicts?

A Salesforce org integrates with an external service using a Named Credential. Security requires mutual TLS and that certificate rotation be manageable without code deployments. Which implementation best satisfies this?

A global retailer wants to replicate a subset of Salesforce data into a cloud data lake for analytics. They need near-real-time ingestion, ordered change events per record, and the ability to replay events after downstream outages. Which solution is MOST appropriate?

A composite API-based orchestration occasionally creates duplicate child records when a client retries after a timeout. The client cannot reliably detect whether the prior request succeeded. What is the BEST architectural mitigation to ensure exactly-once behavior at the application level?

An enterprise is assessing an existing integration landscape. Salesforce is connected to multiple systems through point-to-point REST APIs. Incident reviews show repeated outages caused by tight coupling and inconsistent error handling. What should the architect recommend FIRST to reduce systemic integration risk?

A business stakeholder says, "Customer address changes must show up in our billing system quickly, but it’s okay if it takes a few minutes." The billing system cannot accept inbound calls and only supports pulling data through an API. Which integration requirement best fits this need?

A company must integrate Salesforce with an external claims system. The integration must ensure exactly-once creation of a Claim record in Salesforce even if the external system retries due to timeouts. What is the best requirement to capture for the Salesforce-side behavior?

Salesforce must publish customer events that several downstream systems consume. Some consumers need the full customer profile, while others need only a subset and different transformation rules. The architect wants to minimize coupling between publishers and consumers. Which design is best?

An architect is designing error handling for an asynchronous integration from Salesforce to an external fulfillment system. Requirements: no data loss, ability to replay failed messages, and clear operations ownership. Which approach best meets these requirements?

A team is building a bi-directional integration where both Salesforce and an ERP can update Account fields. Conflicts occur when updates happen close together. What is the best architectural strategy to manage data consistency?

An integration built with Apex callouts intermittently fails with "Too many callouts" errors during high-volume data loads. The process currently makes one callout per record inside a trigger. What is the best fix?

A security review finds that an integration user has "Modify All Data" and broad object permissions to simplify multiple integrations. Auditors require least privilege and traceability per integration. What should the architect recommend?

A global org must integrate Salesforce with an on-prem system that can only be reached through a corporate network. The integration must support outbound callouts from Salesforce without exposing the on-prem endpoint to the public internet. Which solution best fits?

Operations reports that an event-driven integration occasionally processes messages out of order, leading to incorrect downstream state (e.g., an "OrderShipped" event processed before "OrderCreated"). The architect wants to preserve logical ordering per order while maintaining high throughput across many orders. What is the best approach?

A company uses Salesforce for CRM and an external ERP for order fulfillment. A new integration must ensure that Sales reps can see whether an Order is "Shippable" in near real time. The ERP can publish events but cannot be called synchronously during peak hours. What is the recommended integration pattern?

An integration team needs to decide which system should be the system of record for customer email addresses. Salesforce users frequently update emails, but an MDM system enforces validation rules and uniqueness across brands. What is the best first step to translate this business need into integration requirements?

A company must integrate Salesforce with multiple downstream systems. Security requires short-lived access tokens and no long-term storage of user credentials in the middleware. Which approach best meets this requirement for server-to-server integrations?

A Salesforce org receives a high volume of updates from an external system. The external system frequently sends the same update multiple times due to retries. Which design choice best prevents duplicate processing in Salesforce?

A middleware platform calls Salesforce APIs from multiple internal apps. The security team wants each app’s access to be independently revocable and auditable without impacting other apps. What is the best approach?

Salesforce must expose a service for partners to create Cases. Partners have varying capabilities and the company wants to evolve the payload over time without breaking existing integrations. Which approach best supports versioning and backward compatibility?

An external billing system needs a nightly extract of all Accounts changed in the last 24 hours, including related Contacts. The extract must be reliable and resumable if the job fails halfway. Which solution is most appropriate?

An integration occasionally fails with "UNABLE_TO_LOCK_ROW" errors when updating the same Account from multiple concurrent processes (UI edits, middleware updates, and a nightly batch). What is the best architectural mitigation?

A global company needs near real-time replication of customer updates from Salesforce to a data lake. The data lake requires exactly-once processing and the integration must survive subscriber downtime without losing events. Which approach best fits these requirements?

A company runs a hub-and-spoke integration where Salesforce publishes events to middleware, which enriches them and updates several back-end systems. During an incident, the middleware was down for 3 hours. After recovery, systems are inconsistent and there is no definitive way to reconcile what was processed. What design improvement most directly addresses this class of failure going forward?

An Integration Architect is assessing a new CRM-to-ERP integration. The ERP exposes an API but has strict constraints: it supports only one concurrent session per integration user and enforces a low daily API call limit. What is the BEST first step when evaluating the current system landscape?

A business stakeholder states: "When a customer changes their shipping address, all systems must immediately reflect it." The architect learns some downstream systems only accept address updates in nightly batch windows. What should the architect do NEXT to evaluate business needs?

Salesforce publishes an event to an external order system. The order system is occasionally offline for several hours. The business requires that no order events are lost and that events are delivered once the order system is back online. Which integration requirement should the architect document to translate needs into integration requirements?

A company needs to integrate Salesforce with a partner API that requires mutual TLS (mTLS) and also requires rotating client certificates regularly. The integration must be centrally managed and support multiple Salesforce orgs. Which approach is MOST appropriate?

An enterprise wants a canonical customer model shared across Salesforce and three downstream systems. Each system currently uses different identifiers and slightly different field definitions. What is the BEST architectural approach to reduce coupling and support long-term change?

A Salesforce org receives product updates from an external PIM. Updates must be applied in bulk and may include 200,000 records nightly. The integration must avoid Salesforce locking issues and stay within platform limits. Which solution is MOST appropriate?

A team uses an asynchronous event-driven integration where downstream systems subscribe to customer update events. They start seeing duplicate updates processed downstream, causing occasional data churn. What is the BEST corrective action?

An external system calls a Salesforce REST endpoint to create Cases. During peak load, the caller receives intermittent HTTP 503 errors. Salesforce debug logs show "CPU time limit exceeded" in an Apex REST service. Which change is MOST appropriate to improve reliability?

A global integration uses a middleware layer to upsert Accounts into Salesforce. Occasionally, the middleware retries after a timeout and creates duplicate Accounts because the unique identifier is not consistently used. The business requires that retries never create duplicates. What is the BEST design?

A regulated company must implement an integration that exchanges sensitive customer data between Salesforce and an on-prem system. Requirements include: end-to-end encryption in transit, strict outbound endpoint control, and the ability to prove which integration user accessed which data. Which combination BEST satisfies these requirements?