Zendesk Salesforce integration: Setup guide, key challenges & solutions

Support stacks are rarely as connected as people assume. In MuleSoft's 2025 Connectivity Benchmark, conducted with Vanson Bourne and Deloitte Digital across more than 1,050 IT leaders, the average organization ran 897 applications. It integrated only 29% of them, while 90% said data silos were creating business problems.

‍

A support-to-CRM connection is one of the first silos worth closing, but only if you know what it impacts. This guide covers what the Zendesk Salesforce integration syncs, how to set it up, and when the native app stops being enough. 

‍

Since we run these integrations for clients as part of our human + AI support model, we will also talk about common pitfalls in production. 

What the Zendesk Salesforce integration actually syncs

Most teams describe the Zendesk Salesforce integration as one two-way sync between support and CRM. That assumption causes a good share of the setup problems we run into. 

‍

In reality, the native integration, built and maintained by Zendesk, bundles 4 capabilities that people tend to group as "sync." Three of them are genuine data syncs with a fixed direction. The fourth pair are live views into the other system. See the picture below for a breakdown.

‍

What does that mean for your data?

What really separates the 4 functions is the direction data travels. Here’s how each of the flows works once it’s live:

‍

• Customer data sync: Salesforce accounts and contacts write into Zendesk as organizations and users, and nothing on the Zendesk side comes back.
  
  
• Ticket sync: New and updated Zendesk tickets become Salesforce cases, but a case edited in Salesforce isn’t reflected in any ticket.
  
  
• Ticket view: Authorized Salesforce users can open and even edit tickets through it, since they're acting on the original record.
  
  
• The support app: It shows chosen Salesforce fields in the Zendesk sidebar.

We encourage you to check out Zendesk's own documentation for a better understanding of the integration process. Remember that the data sync is, by design, one-directional. So, when your agent edits a Zendesk ticket from inside Salesforce, no data crosses between systems. They're basically changing the one record that already exists, through a different screen.

Access and licensing you need first

Since most of the setup friction is permission-related, it makes sense to sort out all the major questions from the get-go.

‍

You’ll need:

‍

• A Zendesk Support administrator and a Salesforce administrator, or a Salesforce user with the documented profile permissions.
• A full Salesforce user license that can sign in interactively, since API-only and Integration-user licenses aren't supported for the connecting user.
• Salesforce API access. Performance, Unlimited, Enterprise, and Developer subscriptions all include it within the package. 
• Streaming API permissions if you plan to sync account, contact, or lead data.

Note: As of 2026, Zendesk supports connecting multiple Salesforce organizations and Zendesk accounts, replacing the old one-to-one restriction. If you tried this a couple of years ago and gave up, the constraint you hit may be gone.

{{cta}}

How to configure the Zendesk Salesforce integration setup

The first thing to know is that the Zendesk Salesforce integration setup runs from the Zendesk Admin Center. Second, Zendesk recommends completing the process in sandboxes first to ensure production permissions and record IDs match the desired organization. 

‍

Third, here is the setup process, with a step-by-step breakdown below. 

Step 1. Check the Salesforce session settings

In Salesforce:

‍

• Go to Setup → Security → Session Settings.
• Clear "Lock sessions to the IP address from which they originated." 
• Alternatively, set up Zendesk's allowed-IP alternative. 

What can go wrong: A locked session lets the first OAuth connection succeed, then drops it when the session expires, which reads like an intermittent sync bug instead of a wrong policy setting. 

‍

If a working connection starts dropping, check the setting. 

Step 2. Connect Zendesk to the Salesforce system

In Admin Center:

‍

• Open Apps and integrations → Integrations → Salesforce
• Add a connection
• Pick Production or Sandbox
• Accept the terms and authenticate with your approved Salesforce user. 

What can go wrong: Authentication fails when the connecting user is API-only instead of a full user who can sign in interactively. And selecting the wrong instance misroutes the whole setup. 

‍

Label each connection clearly before you switch anything on, so you don’t guess what’s what.

Step 3. Install and configure the Salesforce Support app

Select the connection, open the Support app, and turn it on. From there:

‍

• Pick which Salesforce objects and fields agents should see.
• Set the refresh behavior.
• Preview the layout and save.

What can go wrong: Even after the app installs cleanly, agents can see an empty sidebar when the integration user lacks object-/field-level access to the Salesforce objects you mapped.

‍

Give the integration user read access to every object and field you mapped, then reload the ticket.

Step 4. Set up Ticket View in Salesforce

In the connection's Ticket View tab:

‍

• Install the managed package
• Enable Ticket View
• Define your matching criteria
• And grant access to the right Salesforce profiles. 

Then enable PKCE in the connected-app settings and pre-authorize your admin-approved users.

‍

What can go wrong: Users without the managed package or profile access see no tickets at all. Matching, non-unique criteria are the bigger risk because they can display one customer's tickets on another customer's record.

‍

Grant the package and the right profiles to the users who need them, and match on a unique field such as an external ID rather than name or email.

Note: Some older guides may tell you to disable PKCE. Zendesk's current documentation, however, says to enable it in the Salesforce connected-app settings so that assigned, non-admin users can see Ticket View. 

Optional: add true record replication

Having taken the 4 steps above, you will have visibility and ticket-to-case synchronization. However, at this stage they won’t copy records between Zendesk and Salesforce. 

‍

If you need that, set up the 2 syncs on top, configuring each one on its own:

‍

• Data sync copies Salesforce accounts and contacts into Zendesk as organizations and users. Map the fields for each pair, and match on a unique key such as an external ID.
  
  
• Ticket sync copies Zendesk tickets into Salesforce as cases. It needs a dedicated Salesforce case record type, the managed package, field mappings, and the Zendesk triggers that fire the sync.

Keep that case record type separate from your existing case processes, so support tickets don't get mixed into sales workflows. If the mapping and triggers start to look like a whole project, look into Zendesk integration services to set it up for you and test it inside your existing org.

Zendesk Salesforce integration problems and how to fix them

As you may have grasped, most Zendesk Salesforce integration problems aren't outages. More often than not, they are configuration issues that linger after integration. 

‍

In the table below, we’ve collected the major issues we’ve highlighted along with their usual cause and fix. 

Common Zendesk–Salesforce integration problems and fixes

CRM data is messy to start with. Validity's 2025 State of CRM Data Management report found 76% of users say less than half their CRM data is accurate, even though 90% call it the cornerstone of their operations.

‍

And most of these problems come back to the match key. Email and account name are easy to match on, but they change often, which makes them a weak key. Where the app supports it, match on a stable external ID instead. 

HubSpot Zendesk integration: two apps, two different jobs

Where Salesforce connects through a single native app, HubSpot splits the Zendesk connection across 2 of them. One is built by Zendesk, the other by HubSpot, and they do different jobs. That's what makes the HubSpot Zendesk integration trickier than the Salesforce one looks. 

Zendesk-built app vs HubSpot-built Data Sync

{{cta}}

Setting up the Zendesk-built timeline app

Install the app built by Zendesk from the HubSpot App Marketplace using a Super Admin or a user with App Marketplace permissions. Authorize both sides. 

‍

The setup creates 3 Zendesk triggers and 1 webhook for ticket created, ticket solved, and CSAT submitted events. Test it by opening a HubSpot contact, filtering the activity feed to Integrations → Zendesk, and running a new ticket through. Confirm the activity lands on the contact timeline.

Setting up the Zendesk HubSpot data sync

After installing HubSpot-built Data Sync, go to: 

‍

• Settings → Integrations → Connected Apps → Zendesk → Set up your sync. 
• Independently configure Users ↔ Contacts and Organizations ↔ Companies.
• Name the authoritative system for conflicts on each pair.

Review the default mappings; custom field mappings require Data Hub Starter or higher. Run a limited pilot before the first historical sync, which HubSpot says normally completes within 10 minutes after initial indexing.

HubSpot Zendesk integration limitations

The HubSpot Zendesk integration limitations mostly regard direction and customization:

‍

• Contact, not company. The Zendesk-built app places ticket activity on the HubSpot contact timeline, and current Zendesk documentation says that's the only place it displays.
• No general two-way ticket sync. The Zendesk-built app sends discrete created, solved, and CSAT events.HubSpot Data Sync can sync contacts and companies two-way, but tickets still sync one way only.
• No custom ticket fields in the Zendesk-built app. Its payload and displayed ticket fields can't be customized.
• Archived tickets are excluded. HubSpot's marketplace listing states that archived Zendesk tickets don't sync through Data Sync.

Neither HubSpot app gives you field-by-field, two-way ticket sync. If a ticket's status has to look identical in both systems at all times, you're past what these apps do, and into middleware or a custom build.

Freshdesk CRM integration: Salesforce and HubSpot options

If you haven't settled on a help desk yet, our Freshdesk vs Zendesk comparison is a better place to start than any of these integrations. But if you are sure you will be integrating Freshdesk as your CRM of choice, then you should know the following.

Freshdesk Salesforce integration

Freshworks' current Salesforce Connector App has 2 parts: 

‍

• a Salesforce data widget inside the Freshdesk inbox
• and data sync powered by the paid Connector Apps add-on. 

For contacts and accounts, you can sync in either direction or both ways, once you've mapped the fields and set a primary key. The add-on is documented at $80 per 5,000 tasks and needs an active Freshworks subscription. 

‍

An older "Freshdesk for Salesforce" package was deprecated, so start from the current Connector App rather than following a guide that names the retired one.

Freshdesk HubSpot integration

This can mean 3 of the key applications:

‍

• A free Freshworks-built app that syncs HubSpot marketing contacts into Freshdesk contacts, and nothing more
• HubSpot's own Data Sync app for companies, contacts, and tickets, which advertises two-way record updates and needs Data Hub Starter or Professional for custom field mappings
• Freshworks' own HubSpot connector, documented for Freshchat and Freshdesk Omni specifically, not every Freshdesk plan.

So before you commit, check which company built the app and confirm in your Freshdesk admin console that it supports your plan. Two-way sync isn't available in all of them.

Native app, middleware, or custom API: how to choose

Those are the 3 CRM integration methods for any help desk, whether you're on Zendesk, a HubSpot-based stack, or one of the newer AI help desk platforms that bundle customer service chatbots and CRM sync by default. The right method depends less on company size than on who owns the data and how you handle exceptions. Here's how the trade-offs fall for a typical helpdesk CRM integration.

‍

Take the first of these that matches your situation:

‍

• The native app's official direction and objects already fit your workflow → start native.
• You need special orchestration, like branching, enrichment, or two-way field sync → use middleware.
• You have proprietary objects, strict latency or compliance rules, or scale that justifies owning the whole lifecycle → build custom.

Even "low-code" routes carry dependencies. Salesforce's own 2026 Zendesk connector for Flow, for instance, needs the MuleSoft for Flow add-on in some scenarios.

Read our rundown of Zendesk alternatives and the broader best help desk software comparison for a faster breakdown of all major software options. 

Your CRM integration checklist before launch

Run this CRM integration checklist before you send anything to production:

‍

• Name an accountable owner for each system and for the integration after launch.
• Confirm admin roles and least-privilege access on both sides.
• Document the source of truth and sync direction for every object and field.
• Export the approved field-mapping sheet and the required-field list.
• Define unique matching keys and dedupe rules before the first sync.
• Test create, update, close, reassignment, and permission-revocation on representative records, including duplicates and special characters.
• Record API limits, alert thresholds, and who owns retries.
• Prepare a rollback plan: disable triggers, preserve IDs, restore affected records.
• Schedule post-launch checks at 24 hours, 7 days, and 30 days.

‍

Note: Data ownership belongs in your launch criteria too. The matching keys and conflict rules you define here dictate whether the sync holds up once real volume hits it. Give agents a documented knowledge base for human support team alongside it, so the CRM context the integration surfaces turns into consistent answers.

When the integration isn't the bottleneck

A clean integration doesn't fix an unclear process, and no amount of customer service automation closes a gap that's really about ownership. We've connected stacks that worked perfectly and still watched service slow down, because the real constraint was elsewhere: undefined access roles, weak escalation rules, inconsistent data handling, or a queue with too few people in it.

‍

This shows up most at the handoff. A ticket can sync into Salesforce as a case in under a minute and still sit untouched for a day because no one owns this specific escalation path.

‍

So if you bring in an outsourced customer support CRM team to work inside your existing Zendesk and CRM, the operational layer deserves as much attention as the sync config. Everhelp's hybrid AI and human support model runs inside your current tools, with an AI chatbot resolving routine tickets and trained agents handling complex cases.

‍

The payoff is measurable: one Everhelp client moved from a 200-hour wait to a 4-minute response once we rebuilt the operation around the handoffs rather than the tooling. 

‍

Read more on how to set up operations around AI integration in our AI in customer service handbook.

{{cta}}