Replace Salesforce to Salesforce Before Its Spring ’27 Retirement

Select a cross-org sharing architecture that fits your use case ahead of the Spring ’27 Salesforce to Salesforce retirement.
If you have inherited a mature org that shares leads and opportunities with several partner orgs, this is the moment to check whether that sharing relies on Salesforce to Salesforce (S2S), a declarative feature that shares records between two Salesforce orgs. When Salesforce retires S2S with the Spring ‘27 release, you will need to implement a different data integration strategy.
Salesforce offers several capabilities for sharing data, but no single one covers every case in the exact same way that S2S did. This post sorts the current options into two paths: replace the connection with a partner-facing capability, or build the integration with a platform that fits. If you are designing cross-org sharing from scratch, use it to compare paths before you commit.
Understand what S2S accomplished
S2S is a declarative, point-to-point integration between two Salesforce orgs. It shares records between them with no code and no middleware required. It includes these features:
- Bi-directional record sync: The publishing org chooses which objects and fields to share, and the subscribing org chooses what to accept and maps it to its own fields. Forwarded records go into a queue and are sent asynchronously, so changes do not appear in the connected org immediately. Salesforce retries failed sends, and unresolved errors are logged in the Connection History.
- Authentication by mutual invitation: One org invites the other and an admin in the second org accepts, so no client credentials or secrets are created or rotated. Changes from the connected org arrive under a Connection User.
Some replacements go beyond what S2S offers, such as role-based partner access in Partner Cloud and identity resolution in Data Cloud One. Replacing S2S is a chance to close gaps you have today.
Assess how you use S2S today
If you need to update your org-to-org integration strategy because you have S2S in place, start by assessing how your organization uses S2S. If you are starting a new build, skip ahead to the partner scenario and platform options below.
Begin by creating an inventory of every connection in each org. Record what is shared, how records are forwarded and accepted, and who owns the connection. Also investigate what depends on the connections, such as automation, assignment rules, and logic that reacts to updates from the Connection User. Capture these dependencies as a baseline for your design.
Next, plan to retire any connections that are no longer needed, and then use the list of the remaining connections to gather the business requirements for each use case that they support.
Review the S2S retirement timeline
See the phased retirement schedule, Salesforce recommended alternatives by use case, and guidance on testing in a sandbox before you cut over.
Replace S2S with Experience Cloud or Partner Connect in a partner scenario
If the other org belongs to a partner, you have two options: Experience Cloud or Partner Connect, which requires Partner Cloud. The choice depends on whether the partner needs their own copy of the data.
Use Experience Cloud when partners can work in your org
The requirement is shared visibility into leads, opportunities, or cases, where partners update progress and you maintain one authoritative view. Co-selling is a common S2S use case: you forward records to a partner’s org, the partner works on its own copy, and updates sync back. Experience Cloud gives partners a site where they log in and work on the records you share, so there is no copy to sync.
Set Default External Access to Private, and grant access through partner roles and sharing rules.
The trade-off is that partners work only in your system and have no copy of their own. To mitigate that, give them reports on their own pipeline or cases.
Add Partner Cloud when you manage a channel partner program
Partner Cloud builds on the Experience Cloud site described above. It adds onboarding and training, tiers and business plans, and channel revenue management such as incentives and market development funds.
The sharing model and partner access design carry over, so you extend them instead of rebuilding. The cost is license spend, since Partner Cloud requires Enterprise Edition or greater, plus program design work such as tier and incentive rules.
Use Partner Connect when partners need their own copy of the data
If a partner must keep its own copy of the data, consider using Partner Connect (Beta), an add-on for Partner Cloud. Partner users export approved records from your site into their own org, and changes to the synced fields flow between the two copies.
As of the Winter ’27 release, Partner Connect supports accounts, contacts, leads, opportunities, and cases. It also supports only some field types, for both standard and custom fields. For example, it does not support formula, multiselect, or most lookup reference fields.
If a beta feature is not acceptable in production, MuleSoft can replicate records between your org and the partner’s, so each side keeps its own copy. Build a custom integration only if no option fits.
Build the integration using Salesforce Connect, Data 360, or MuleSoft
If a partner-facing capability doesn’t fit, build the integration on a platform. Choose by how much data has to move. Start with the option that moves the least, and go further down the list only when a requirement forces you to. Suppose the requirement is a shared view of customers and open work across orgs. Teams need to be able to route leads, opportunities, and cases without duplicating data.
Work in place with Salesforce Connect
When orgs need to see or update records without copying them, use the Salesforce Connect cross-org adapter. The data stays in the source org and appears on demand as external objects, which can be read-only or writable. It requires a Salesforce Connect add-on license, and each call counts against the source org’s API limits. A slow or unavailable source org fails the page with no automatic retry. The cross-org adapter is best for on-demand, small-volume operations due to timeout and payload size limits.
Salesforce Connect Cross-Org Adapter Developer Guide
Set up Salesforce Connect to access data in another org with the Cross-Org Adapter. Provide users with seamless access to data in your other Salesforce orgs so that they have a complete view of the business.
Share data for insight and activation with Data 360
Data 360 fits when orgs need a shared view for insight and activation.
Data Cloud One gives you a unified customer view in a single Data 360 instance. A home org hosts it, and the other orgs connect as companion orgs, which centralizes governance and identity resolution. The trade-off is dependency and cost: companion orgs can’t ingest or unify data, so they rely on the home org, and their activity draws credits from it. Before you connect orgs, plan your data spaces, identity resolution rules, and how you activate unified data back into each org.
If an org must keep its own Data 360 instance for data residency or autonomy, use Data 360 to Data 360 data sharing instead. Each org shares selected objects, including across regions, and the receiving org gets them as data lake objects for insight and activation, not for editing source records. The trade-offs are cost and governance: sharing consumes credits in both orgs, the receiving org must reapply governance tags, and unification across instances isn’t automatic. Share only the objects each use case needs. To choose between the two, see the Data 360 provisioning decision guide.
Data sharing doesn’t create or update CRM records, so for a process handoff you build the handoff yourself. For example, a data action or flow can call custom Apex to write the result to a CRM record, which adds build and maintenance cost.
Replicate records when each org needs its own copy
Use MuleSoft for Flow when each org must edit its own copy, and an admin team will own the integration. Flows call prebuilt connectors in Flow Builder, so you build and change the sync declaratively, but conflict handling and error alerts remain yours to design. Credit use grows with the records you read, create, update, or delete, so keep volume modest. MuleSoft for Flow is licensed through MuleSoft packages, so check what your package includes against your contract. If one simple flow is your only need, the package cost may be hard to justify, so consider building a custom integration.
Use MuleSoft Anypoint Platform when your organization already runs it, or when a handoff involves more than two systems, needs data transformation, or must be monitored and governed in one place. Each org connects to an integration layer, so a change to a mapping or credential happens in one place. Anypoint lets your team design and operate retry, error handling, and monitoring in the integration layer. If you already run Anypoint, you can reuse your subscription, developers, and deployment model. If you don’t, those become new costs.
Build a custom solution when no option fits
Build a custom solution when you have evaluated the options above against your requirements and none meets the needs of your orgs. A missing feature, a volume or latency limit, or a license cost the use case cannot justify can each be a valid reason.
Common building blocks include Apex and REST API calls for steps a user triggers, Change Data Capture or platform events for changes published from the source, and Bulk API 2.0 for scheduled loads.
A custom build means you are responsible for what the products would have handled. Design for duplicate delivery so a retry does not create a second record, plan how you backfill after an outage, handle deletes, and manage credentials between the orgs. Check callout limits and event allocations to confirm the design handles your peak volume. Plan for error handling, monitoring, and long-term maintenance, with a named owner for as long as the orgs share data.
Plan a controlled implementation
Decide whether your use cases can move in phases or need one coordinated cutover. Then work through these areas:
- Work with the other org’s owner: If the other org belongs to a partner or customer, share your timeline and ownership rules now, including who owns each field and who responds when a transfer fails.
- Plan for automation dependencies: Identify every flow, trigger, report, and integration that touches the data you are moving, and rewrite or remove what no longer applies. Flows, triggers, and validation rules also run on records the new integration writes, so test how each behaves.
- Plan for duplicates and failures: Decide how you will detect and prevent duplicate records, such as by matching on a shared identifier. Then test in a sandbox with pass criteria set in advance: peak volume, child records arriving before their parents, a validation rule rejecting a record, and the other org being unavailable. The Reliability pillar from Salesforce Well-Architected applies here: decide how each integration fails, recovers, and alerts before you decide how it syncs.
- Handle S2S specifics: If you are replacing S2S, look up your instance’s Spring ‘27 upgrade date on Salesforce Trust and finish production cutover several weeks earlier. Run your sandbox tests where S2S is not configured, since running both is likely to create duplicates.











