5 More Things Worth Knowing About the September Gainsight CS Release + Webinar Recording

Thanks to everyone who joined us for our September release webinar! We covered a lot during the session, so rather than recap every update, I wanted to share a few things that I think are worth trying first.

If you missed the session or want to revisit any of the updates, just click play.

1. Relationship Resolution uses more than one signal to find the right Relationship

Relationship Resolution looks at two things at the same time.

First, it compares the people on the interaction against Relationship Person data and configured User data. We also use AI to identify the product discussed in the conversation based on your Organization Profile.

We then use those results to determine the product level Relationship. If neither method returns a match, we log the activity to the Company instead.

That makes the underlying configuration important. We recommend that admins keep Relationship Person data accurate, choose User fields that distinguish one Relationship from another, configure product identification fields with unique values, and include the names customers actually use for each product in the Organization Profile.

One detail worth checking: if customers commonly use abbreviations, former product names, or product nicknames, you can add those terms to the Organization Profile too.

2. Stakeholder signals can do more than add context in 360s

For CS + Staircase customers, you can now bring four stakeholder signals into in P360 or aggregated in C360:

• Stakeholder Not Engaged: Indicates whether the stakeholder is not currently engaged or active in communication
• No Reach Out To Stakeholder: Indicates whether no outbound communication or reach out has been made to the stakeholder
• Stakeholder's Title Has Changed: Indicates whether the stakeholder's official job title has changed or been updated
• Account Personnel Changed: Indicates whether the account personnel associated with the account have changed

The part I don't want you to miss: you can also use these signals in Rules and Programs. For example, if the AI detects that a stakeholder changed titles, Journey Orchestrator can automatically send an email congratulating them.

3. You can combine Staircase and Timeline context in the same question across AI chat surfaces

Ask Staircase now works in Copilot, the Gainsight AI Agent in Slack, and Gainsight CS MCP for joint CS + Staircase customers, which means you can bring communication intelligence and Timeline activity context into the same account question.

Note: In Copilot, you can select Timeline, Ask Staircase, or both as sources. If you use both, Copilot can bring those sources together in one response.

So instead of only asking, “What are the risks for Acme?”, you can ask, “Summarize Acme using Staircase and Timeline” to get an answer grounded in more customer context.

One limitation to remember: Ask Staircase supports qualitative, account- specific questions. It does not support global questions across accounts or quantitative questions such as meeting counts.

4. Evaluate cloning in JO keeps most of your logic, but not every condition

The September release now allows admins to clone a configured Evaluate element inside a draft Journey Orchestrator Program.

When this happens, Gainsight copies the existing Evaluate configuration into the new element, so you can reuse similar branching logic elsewhere in the Program.

But these is one important exception: If a condition references a survey answer or a previous step, we clear that value in the clone. So, if this is the case, make sure to connect the cloned element to its new path first and then reconfigure those conditions before you publish.

5. You can now tell a lot about how your teams are using AI

New CS MCP Logs give you a detailed record of AI activity across Copilot, the Gainsight AI Agent in Slack, and Gainsight CS MCP.

And there’s quite a bit you can dig into.

Each row represents a single AI tool call. We capture the exact user query, the tool the AI called, who triggered it, whether it performed a read or write action, where the request came from, when it ran, and whether it succeeded or returned an error.

That means you can build reports to answer much more specific questions, such as:

• What questions are people actually asking?
• How much activity involves reading data versus taking an action?
• Does usage come from Copilot, Slack, ChatGPT, Claude, or another connected client?
• Which tool calls succeed, and where do errors occur?

There’s another useful connection in this release, too. We added an Update Source field to the Call to Action and Success Plans that identifies whether Copilot or Gainsight CS MCP created or last updated the record.

Basically, CS MCP Logs can show you the AI activity itself, while the Update Source field gives you additional visibility into where AI touches CTAs and Success Plans.

There’s a lot more in the release, so if something came up during the webinar that you want us to unpack further, let us know. You can also dive into the full release notes here.