How to Switch from Intercom: A Practical Migration Guide

Switching off Intercom comes down to four steps: export your data, stand up the new tool and import into it, swap the Messenger snippet on your site, then run both tools in parallel for a short window before you cancel. The export and the widget swap are quick. The part that actually takes time is rebuilding the automations and macros Intercom will not hand you as portable data. Plan for that and the rest is mechanical.

This guide is deliberately honest about what migrates cleanly and what does not, because a migration that surprises you halfway through is a migration that stalls. We will use SupportWire as the worked example of the destination, since that is the tool we know cell by cell, but the steps are the same wherever you land.

Jump to a section:

• The four steps at a glance
• Step 1: Export your Intercom data
• Step 2: Stand up the new tool and import
• Step 3: Swap the Messenger snippet
• Step 4: Run parallel, then cancel
• A realistic timeline
• The migration pitfalls checklist
• About the Salesforce acquisition

---

The four steps at a glance#

Every helpdesk migration, regardless of the tool you are moving to, is the same four moves:

1. Export your historical records out of Intercom (conversations, contacts, articles).
2. Import and rebuild in the new tool: load the records, then recreate the configuration that does not travel.
3. Swap the widget so new conversations start landing in the new tool instead of Intercom.
4. Run both in parallel long enough to be confident, then cancel Intercom on its own timetable.

The order matters. You do not cancel first and scramble. You get the new tool fully live, prove it in parallel, and only then wind Intercom down. Intercom gives no early-exit refund, so there is zero incentive to rush the cutover. Do it calmly.

The full step-by-step version of this, with the bill comparison built in, lives on the migration playbook. Keep it open next to this guide.

---

Step 1: Export your Intercom data#

This is where most of the honest surprises live, so it is worth being precise. Intercom splits cleanly into three kinds of data, and they export very differently.

Contacts and users export cleanly. In the Contacts tab, select the records and click Export, and you get a CSV with contact ID, name, business name, phone, email, creation date, last activity, and tags. Users, leads, and companies work the same way through More, then Export. This is the easy part. See Intercom's export users and companies guide for the exact clicks.

Conversations are the trap. There is a CSV export under Reports, then Dataset export, and it is tempting to assume that is your conversation history. It is not. The CSV is metadata only: it carries the fields about each conversation, not the actual messages inside it. It also caps at 10,000 rows for an instant browser download, with larger pulls emailed to you. If you want the real transcripts, you have two options: the per-conversation PDF or text export, which does include message content and can be run in batches of a couple hundred at a time, or the Conversations API for full-fidelity export beyond the CSV limits. The API needs an admin token and a bit of developer time, and it pulls history in 90-day windows per request, so a multi-year backfill is several sequential calls. Intercom documents the mechanics in its export conversations guide. One more thing to know: internal notes are never included in any transcript export, and even the S3 JSON export has transcript gaps, so do not count on private agent notes coming across.

One practical note on scope: Intercom keeps historical data available up to about two years back, and exporting the CSV requires the "Can access Chart drill-in" and "Can export CSV" permissions, so make sure whoever runs the export has them before go time. It is a small thing that stalls a lot of migrations on day one.

Help center articles have no bulk-export button. As of writing, Intercom's UI has no one-click way to download all your articles. The supported route is the Articles API: list every article to get the IDs, then retrieve each one for its full content. That means an admin access token and a script, or a third-party migration app if you would rather not touch code. Budget for it. A knowledge base of any size is not a five-minute copy-paste.

Macros, automations, triggers, and routing rules do not export at all. This is the single most important line in this guide. Your saved replies, workflow automations, assignment rules, SLA policies, and bot flows are not data you can port. Intercom's own helpdesk migration guide is candid about this: only historical records (conversations, contacts, tags, articles) migrate cleanly. Everything in the "operational configuration" bucket has to be rebuilt in the new platform by hand. Before you cancel anything, take screenshots or copy the text of every macro and every automation you rely on. That backup is your rebuild checklist for Step 2.

---

Step 2: Stand up the new tool and import#

Now you load what you exported and rebuild what you could not. Here is where the destination tool starts to matter, so this is the SupportWire worked example. The moves generalize to any tool you pick.

Import the records. Conversations and contacts go in first. In SupportWire this is a guided import rather than a raw CSV wrestling match, and the point is to get your history searchable in the new inbox so agents are not flipping back to a cancelled account to look things up.

Load your help content into the Knowledge Store. This is the part that pays off fastest. SupportWire's AI agent, Kal, answers customers by reading from the Knowledge Store, a repository of sources you point it at: your help center articles, docs, past resolved conversations, whatever you feed it. The articles you pulled through Intercom's API in Step 1 become inputs here. Once they are in, Kal can start resolving the same repetitive questions your Intercom setup handled, without you rebuilding a single bot flow by hand.

Rebuild the configuration you screenshotted. Recreate the macros as saved replies, the automations as their new-tool equivalents, the routing rules, the SLAs. This is manual everywhere, not a SupportWire quirk, and it is the honest bulk of the timeline. The upside is that a migration is a good moment to leave the stale ones behind rather than carrying a decade of cruft into a clean system.

On cost, the reason a lot of teams start this evaluation is the bill. SupportWire is $19 per seat per month with 50 AI resolutions included, then $0.49 for each resolution beyond that. If you want to see your own Intercom bill next to that number, the Intercom pricing calculator runs it on your seats and volume, and the full feature-by-feature breakdown is on Intercom vs SupportWire.

---

Step 3: Swap the Messenger snippet#

This step is mechanical and fast, and it is the moment new conversations start flowing to the new tool. Intercom does not document the widget swap, because from their side it is out of scope, so here is the general practice.

The Intercom Messenger is a JavaScript snippet living in your site's global layout, usually in the <head> or just before the closing </body>. Removing it means deleting that snippet and installing the new vendor's widget snippet in the same spot. It is a one-time code deploy, not a data migration, but the timing matters: you want the new widget live and tested before you pull the old one, so there is never a gap in live coverage where a customer opens a chat and nothing answers.

For SupportWire specifically, there is a drop-in Intercom Messenger replacement built for exactly this cutover, so the swap is closer to a straight substitution than a rebuild. Whatever tool you choose, treat this as a coordinated deploy: stage it, test it on a non-production page, then ship it and watch the first live conversations land.

---

Step 4: Run parallel, then cancel#

Do not cancel Intercom the day you flip the widget. Run both tools side by side for a window, a week or two is typical, while new conversations flow into the new tool and open Intercom threads wind down naturally. This is your safety net: if something in the import or the rebuild is off, you catch it while the old system is still reachable.

The parallel window is also when you confirm the import actually landed. Spot-check a handful of imported conversations against Intercom, search for a few contacts by email, and open a couple of help articles to be sure the content came through intact. Better to catch a gap now, while the source is still live, than a month after you cancel.

When you are confident, cancel, and cancel on Intercom's terms rather than fighting them. From Intercom's cancellation policy:

• On a contract: you must give written notice at least 30 days before your renewal date, or the contract auto-renews for another 12 months. Contracted accounts cancel through an Account Manager, not the settings page. Build that 30-day window into your plan from the start.
• Annual (non-contracted): cancel under Billing, then Payment details. It takes effect at the end of the current 12-month period. No auto-renewal after that, but no proration or refund for the unused months either.
• Month-to-month: cancel any time in Settings, effective at the end of the current month.

The through-line: there is no refund for leaving early, so there is no reason to rush the cutover and every reason to time the cancellation date to land right after the new tool has proven itself. Let the paid period you already bought run out while you settle in.

---

A realistic timeline#

Intercom's own migration guidance sets the honest expectations here, and they are worth quoting because they are not a vendor trying to sell you a fast switch.

A simple migration using automated tools finishes in one to two weeks. An enterprise-scale migration runs four to eight weeks. And planning alone accounts for roughly 60% of the total effort, so you should budget one to two weeks of planning no matter your size.

Read that carefully: the export is not the long pole, and neither is the widget swap. The time goes into planning and into rebuilding operational configuration by hand. A ten-person team with a light automation setup and a small help center is a one-to-two-week job. A large org with hundreds of macros, complex routing, and a two-year knowledge base is the four-to-eight-week end. Size your expectations to your configuration, not your headcount.

---

The migration pitfalls checklist#

These are the ways helpdesk migrations go wrong, drawn from Intercom's own list and worth checking off before you start:

• You skipped the data audit. Migrating before you know what you have means discovering missing tickets or contacts after the switch, when it is expensive to fix. Inventory first.
• You migrated the mess. A migration is the one clean moment to drop stale articles, duplicate tags, and dead macros. Carrying them over just reinstalls the cruft in a new house.
• You left notifications on during import. Bulk-importing old conversations with notifications live can flood customers with alerts about issues that were resolved months ago. Mute before you import.
• You trained agents at the last minute. Adoption drops and productivity dips when the team meets the new tool on go-live day. Train during the parallel window, not after.
• You let scope creep in. "While we are migrating, let us also redesign the whole workflow" is how a two-week project becomes a two-month one. Migrate first, improve second.
• You forgot the config does not travel. The single most common surprise. Screenshot every macro and automation in Step 1 so Step 2 has a checklist instead of a memory test.

---

About the Salesforce acquisition#

One reason teams are looking at their options right now: in June 2026, Salesforce signed a definitive agreement to acquire Fin, the company formerly known as Intercom. Worth being precise about what that means. The deal is signed, not closed, expected to complete later in Salesforce's fiscal year, and no pricing change, contract change, or forced migration has been announced. It is a reason to calmly evaluate whether your helpdesk still fits, not a reason to make a rushed decision off a headline. If you switch, switch because a different tool serves you better.

---

Whatever you pick, the steps are the same#

The honest takeaway: this four-step shape holds for any helpdesk you move to. Export the records, stand up and import, swap the widget, run parallel and cancel. The destination decides how much of Step 2 is painful and how much your monthly bill drops, not whether the process works.

If you want the guided version with your Intercom bill compared line by line, start at the migration playbook. If you are still choosing where to land, the best Intercom alternatives for startups walks the field with honest takes on each. And if the number is what pulled you here, the Intercom pricing calculator will tell you in about a minute what leaving actually saves.