Most CS teams are operating two separate records of what a customer has communicated to them. One lives in Zendesk or a similar support platform: written tickets, email threads, internal notes from support agents. The other lives in Gong or Chorus: transcripts and summaries of calls the CSM, AE, or support team had with the account. In practice, neither system knows what the other contains.
The account executive looking at Gong before a renewal call sees the call history. They do not see the support ticket where a user wrote that the product "keeps failing in the exact same place" for the third week running. The CSM working through Zendesk sees the ticket queue. They may not know what the account expressed on the last check-in call unless someone wrote it in a CRM note. Both of them are working from a partial picture.
This is not a process failure specific to any particular team. It is a structural artifact of how these tools are designed. Zendesk is built to manage support workflows. Gong is built to improve sales performance and coaching. Neither is designed to synthesize cross-channel customer sentiment at the account level for customer health purposes.
Why the two channels carry different kinds of signal
Support tickets and call notes capture different modes of customer expression, and those modes contain complementary signal rather than duplicative signal.
Support tickets are filed by users: the people who work in the product, encounter the problems, and write about them from inside the workflow. The language is operational and specific. It reflects the day-to-day experience of using the product rather than the strategic view of the relationship. A ticket filed at 2pm on a Tuesday tells you what broke in the customer's workflow right then.
Call notes capture a different conversation, typically between a customer decision-maker or champion and a CS or account management contact. The language is higher-level and more strategic. Objections and concerns that would never appear in a technical support ticket show up here: budget pressure, organizational changes, evaluation of alternatives, uncertainty about renewal scope. These are relationship-level signals, not workflow-level signals.
When a customer is heading toward a difficult renewal, both signals usually appear before the renewal conversation itself. The workflow frustration accumulates in the ticket record over weeks. The strategic uncertainty appears in call language in a different form and often later in the timeline. Seeing both gives you a more complete picture of where the account actually is.
The practical problem with connecting them
The technical challenge is not the API access. Both Zendesk and Gong have APIs that expose ticket data and call transcripts respectively. The engineering problem is the matching and normalization step.
Zendesk tickets are associated with a requester email address and an organization record. Gong calls are associated with contact records in Salesforce or HubSpot, which may or may not map cleanly to the Zendesk organization record. Organizations that go through a name change, a merger, or that operate under multiple domains will have records that do not match automatically. Before you can do any analysis that spans both channels, you need a reliable account-level identifier that works across all three systems.
The second normalization challenge is time. Tickets have a created timestamp, a status history, and in many systems an SLA resolution timer. Call notes have a call date and a duration. But those timestamps do not account for the difference in how information propagates in the two systems. A concern raised in a call may not appear in any system record for 24 to 48 hours if the CSM summaries in Gong are async. A ticket may reference a problem that was first mentioned on a call two weeks prior. Building a timeline that accurately represents the account's experience requires handling these offsets explicitly.
What combined analysis looks like in practice
Once the matching and normalization layer is in place, the analysis that becomes possible is different in kind from what either system supports alone.
The most straightforward case is corroboration: the same concern appearing in both channels within a short time window. If a user filed a ticket about an integration reliability issue and a decision-maker mentioned "the integration stability concerns us" on a call recorded in Gong two weeks later, that corroboration is meaningful. The concern has moved from the workflow level to the decision-maker level. That is a different risk signal than either event alone.
The more nuanced case is divergence: a period where call sentiment is positive but ticket language is shifting toward frustration, or the reverse. Divergence between what users are experiencing and what decision-makers are saying often indicates either that the champion is insulating the vendor from bad news (a risk factor in itself) or that the user-level problems have not yet surfaced in the relationship conversation. Either pattern has implications for how a CSM should approach the next touchpoint.
Temporal analysis is the third valuable dimension. If an account's ticket language started shifting toward the resigned pattern we describe elsewhere in this blog four weeks ago, and the most recent Gong call had the account champion expressing uncertainty about the renewal scope, the alignment of those two signals on the same account timeline is a stronger indicator than either alone.
The gap this leaves in current tooling
The reason most CS teams are not doing this analysis is not that it is technically impossible. It is that the tooling for it does not exist in a packaged, maintained form. Building a custom integration between Zendesk and Gong that handles the normalization, the account-matching, and the cross-channel analysis requires engineering time that most CS ops teams do not have, and the resulting system requires ongoing maintenance as both APIs change.
Gainsight and Vitally both have some integration with both platforms, but their data models are oriented around health scores rather than cross-channel text analysis. They ingest the data but they do not read the content in a way that surfaces communication pattern shifts at the account level.
This is the specific gap Sturdy addresses. We ingest from both Zendesk and Gong, handle the account-matching layer, and analyze the combined communication record for the language patterns that precede churn. The output is account-level flags that reference the actual tickets and call note passages that triggered them, spanning both channels.
A note on what we are not claiming
It is worth being direct about the limits here. Connecting Zendesk and Gong does not produce a complete picture of customer health. Product usage data matters, especially for the adoption phase of an account. The relationship history, the champion strength, the competitive environment: all of those factors influence renewal outcomes and none of them are captured in the communication record alone.
What the combined communication record does provide is the best available signal for sentiment trajectory at the account level, and that is the signal that is most often missing from existing health models. Most teams have the usage data. Most teams have a relationship history in their CRM. What they are typically missing is a systematic read of what the customer has actually been saying, across both the operational and the strategic channel, in the six to eight weeks before the renewal conversation. That is the gap the Zendesk-to-Gong connection is designed to fill.
