Gong Salesforce Integration Explained and How to Set It Up Right

At its core, the Gong io Salesforce integration does two jobs. First, it pulls CRM data into Gong: per Gong's own FAQ, Gong imports Salesforce data from accounts, opportunities, leads, contacts, and users, which is how Gong knows that a Zoom call with three people belongs to a specific opportunity. Second, it pushes conversation data back into Salesforce: calls and emails appear as activity on the right records, and with the managed package installed, each call becomes a Gong Conversation record that holds the recording link, metadata like date, duration, and participants, Gong analytics including topics, trackers, and interaction stats, and the full transcript when available. Salesforce fields also become usable inside Gong for search filters, call recording rules, and Gong Engage email templates, so the integration is as much about context flowing in as insights flowing out.

There are two distinct pieces people mean when they say Gong Salesforce connector. The first is the base API connection, set up from Gong's admin area: you open Integrations, select Salesforce, click Connect, and authorize with a Salesforce admin login. This connection imports CRM data into Gong, and imported users typically appear within a couple of minutes, but on its own it is largely a read path into Salesforce. The second piece is Gong for Salesforce, the managed package installed from AppExchange, which is what writes rich conversation data into your org: the Gong Conversation object, canvas apps for highlights and transcripts, preset reports and dashboards, and permission sets. Most teams that want Gong data visible inside Salesforce need both: the connector for import and matching, and the package for the in-CRM experience.

Gong's install documentation is specific about what you need before installing the package. You need a Salesforce Enterprise license, tech admin access in Salesforce, API access enabled, and at least three free custom fields: two on the Opportunity object and one on Account. You also need free data storage headroom, since each recorded call creates a parent record plus roughly 25 to 40 child records at about 2KB each. Gong also recommends connecting through a dedicated integration user rather than a personal admin account, because if the person who set up the integration is later deactivated in Salesforce, the integration disconnects with them. One more constraint worth knowing up front: only one instance of Gong for Salesforce, either production or sandbox, can be connected at a time.

Installation itself is a standard AppExchange flow. Find the Gong.io listing on AppExchange, click Get It Now, choose whether to install in production or use the sandbox option to trial it, confirm the installation details, and select Install for All Users. The listing shows a price of 1 dollar, but Gong's docs are explicit that Gong for Salesforce is a free app; the dollar is simply the lowest amount AppExchange allows for a listed app, and you are not billed for installing it. The listing requires Sales Cloud and, as of this writing, carries a 4.99 out of 5 rating across 81 reviews. After the install completes you are not done: the post-installation configuration in the next sections is what actually makes Gong data visible to your users.

On the import side, Gong lets admins choose which standard and custom fields to bring in from accounts, opportunities, leads, contacts, and users via the data import settings. Those fields then power call-to-deal matching, filtering, and recording rules. There are real limits here: Gong's FAQ states it cannot import multi-select picklists, address fields, location fields, encrypted fields, or complex reference fields, so if your team leans on any of those for segmentation you will need a workaround field. Gong also notes it only imports opportunities for users who have data capture enabled, which trips up teams that enable Gong for a subset of reps and then wonder why some pipeline is invisible. Authentication uses OAuth, and Gong states it does not store your Salesforce credentials, only the tokens.

On the export side, Gong logs activity in Salesforce in two ways. It creates tasks and events on the matched records for meetings, conference calls, telephony calls, emails sent through the Gong email composer, Engage tasks, LinkedIn requests and messages, and flow activity, each logged with consistent prefixes and field values so admins can report on them cleanly. Separately, the managed package writes each call to the Gong Conversation object with its recording, metadata, analytics, and transcript, and if call playback is enabled reps can listen to calls directly inside Salesforce. Volume-wise, Gong estimates approximately one API call per recorded call against your Salesforce API limits. Note the deletion behavior too: when a call is deleted in Gong, the corresponding data is deleted in Salesforce as well.

Gong Engage, the outbound engagement layer, rides on the same integration. Emails sent from Engage's composer are logged to Salesforce as activity, Engage tasks and flow steps show up on the matched contact and account records, and Salesforce fields can be used as merge fields inside Engage email templates. The post-installation guide also covers adding Engage Flows as a related list on page layouts so reps can see which flows a contact is enrolled in without leaving Salesforce. If Engage emails are not appearing in Salesforce, the usual first check is whether the sending address is listed under Gong captured emails in Gong's admin settings, since Gong only logs activity for addresses it has been told to capture.

The package ships with preset permission sets that handle most of the access model: Gong Users for end users, Gong Admin for system administrators, and Gong API User for integration users that need contact or lead creation. The Salesforce integration user needs read access to Account, Opportunity, Contact, Lead, Opportunity Stage, Opportunity History, and Publisher, plus Task if you use a dialer integration, and write access to Task and the Gong Conversation object if you want calls and emails pushed. Seeing highlights and transcripts also requires enabling the Gong Canvas app under Manage Connected Apps, setting Permitted Users to admin approved users are pre-authorized, and adding the relevant profiles. In practice, most cannot see Gong data in Salesforce complaints trace back to a missed permission set assignment or an unconfigured canvas app rather than a broken sync.

The honest answer is that the integration is two way but asymmetric. Salesforce to Gong is a broad, continuous import of records and fields; Gong to Salesforce is a narrower, well-defined write path of activity logs and conversation records, and Gong cannot, for example, automatically create contacts in Salesforce through the standard integration. That asymmetry is a feature: automated writes into a CRM should be scoped and auditable. It matters more once teams layer AI on top of call data, drafting opportunity field updates, next steps, and follow-up emails from what was said on the call. Our view at Mindlyft, and the way we built ASTRA, is that those drafted updates should sit in a review queue where a CSM or AE approves, edits, or rejects each one before it touches Salesforce, so the CRM stays trustworthy while the busywork disappears. If you want post-call updates flowing into Salesforce, your ticketing system, and follow-up emails with that approval step built in, we build the first workflow free, then it is 5,995 dollars per month as a GTM engineering subscription at mindlyft.in.