Question 1

Which Intercom objects accept writes from Databricks?

Accepted Answer

Leads, users and companies. A good deal of other Intercom data can be read into the lakehouse, including contacts, tags, conversations, admins, teams and tickets, but those are source only. The writable set is the useful one for this direction anyway, since it holds the attributes an agent actually sees when a conversation opens.

Question 2

What happens if several syncs hit the warehouse at once?

Accepted Answer

They queue and compete, which shows up as variable runtimes rather than obvious failures. Every reverse ETL run is a real query on the SQL warehouse you configured, and Polytomic does not spin up separate compute per sync. Setting a query concurrency limit is the practical guard, since it keeps runs predictable and avoids capacity errors when several destinations refresh on the same schedule. Worth doing before you add the fifth or sixth sync rather than after.

Question 3

How does a user get linked to its company?

Accepted Answer

Through a column in the model. Include the company identifier next to the person and map it, and leads or users are associated with that company in the same sync rather than needing a second pass. Worth doing from the start, because the context that helps an agent is usually company level: which plan the account is on, what it is worth, when it renews.

Question 4

Should the model be allowed to create Intercom users?

Accepted Answer

Usually not, if your product already creates them. Update or create will make anyone missing, update only enriches those already there, and create only adds new ones without touching matches. Since Intercom is typically fed by signup, letting a lakehouse model create people as well tends to produce duplicates from rows that never went through that path. Update only keeps ownership clear and still gets the attributes onto the records that matter.

Question 5

Is this worth it if the product already sets attributes?

Accepted Answer

Often, because the two carry different things. Your application can send what it knows about a session or a plan, but not anything derived from billing history, support volume or a model combining several systems. Reverse ETL fills that gap without asking engineering to build another integration, and keeps the definitions identical to the ones reporting uses rather than close but not quite.

Sync Databricks to Intercom

Sync Databricks to Intercom with Polytomic

Users, leads, companies

Attach people to companies

Concurrency worth watching

Or sync Intercom to Databricks

Trusted by businesses worldwide

Step-by-step integration guide

Connect Databricks in Polytomic

Connect Intercom in Polytomic

Create a sync between Databricks and Intercom

Map fields and enable the sync

Frequently asked questions

The sync and ETL platform that does it all

Powerful data sync

Advanced analytics

Enterprise security

Easy setup

Ready to connect Databricks with Intercom?