Skip to main content

I’m curious as to how teams are leveraging Copilot to support their Customer Success organization.

I would also love to hear from the product team as to what the direction of the feature is and what enhancements we can expect. We have a pilot group of Customer Success Managers who find it very challenging to use because the prompts need to be so specific and compared to other AI out there it doesn’t seem like the most user friendly...

Below is some feedback from our team: 

“I noticed you have to be a bit careful with how you ask it things. It's not as "easy" as Chatgpt or Gemini where it tries to find the closest thing to what you asked "intuitively." For example, when asked "What are the main risks noted for [Company Name]?" it acted as if that account is non-existent. However, when I used the same language Co-Pilot had used in a previous question in the same chat, it worked. Moreover, it was then able to give me suggestions and recommendations to take based on the data it retrieved, which I thought was super useful. I think the actual recommendations could've been better”

“I used it late last week and was trying to see if it would pull data for me in terms of charts we see in widgets, but it looks like it just pulled data from the meeting notes, timeline details, etc. It wasn’t particularly helpful as I knew those details and I was trying to see if it would pull the charts for me...”

“I wonder what's the functionality they've built into copilot so far and if the type of feedback we can give will be taken by GS to enable additional functions. It'd be really helpful if the data it pulls could come from manually entered data like Activity and automated data coming from snowflake, etc.”

Interesting question! We haven’t issued official guidance to our users, but I’m currently running an AI Agent testing phase (very limited-access pilot prior to broader roll-out), and I had Gemini create this doc as a “here’s when to use each” that was based off the result of a deep research query about the different use cases for each tool. The sample prompts are gold.

Doc for reference

Gainsight AI: Copilot vs. Slack Agent - A Quick Guide for CSMs

This guide helps you choose the right Gainsight AI tool for the right task. Both tools are designed to make you more efficient and proactive, but they serve different purposes. Think of them as a complementary pair: one for deep, focused work, and the other for fast, collaborative communication.

The Bottom Line:

  • Use Copilot when you are working inside the Gainsight platform for deep analysis and preparation.
  • Use the AI Agent when you are collaborating inside Slack for quick answers and team alignment.

 

Side-by-Side Comparison

 

 

Gainsight Copilot (In-Platform)

Gainsight AI Agent (In Slack)

Think of it as...

Your deep-dive research analyst and strategic partner.

Your on-the-go messenger for quick facts and team updates.

Core Mission

"More Answers, Less Searching".1 To provide in-depth, synthesized insights for focused work.

"Customer Data, Right Where You Work".1 To provide instant data access for quick collaboration.

Where You Use It

Inside the Gainsight application.2

Inside Slack (DMs).4

Best For

Deep work sessions, strategic planning, and comprehensive meeting preparation.

Quick lookups, real-time status updates, and answering questions from colleagues without context switching.

Primary Audience

You, the CSM, and other Gainsight power users.1

The entire company: CS, Sales, Product, Leadership, and anyone who needs a quick customer fact.1

 

When to Use Each Tool: Practical Scenarios

Use Gainsight Copilot when you need to...

  • Prepare for a major meeting.
  • Example Prompt: "Summarize all value delivered for Customer X in the last 6 months, referencing completed CTAs and positive sentiment from Timeline." 1
  • Perform deep portfolio analysis.
  • Example Prompt: "Identify my top 3 accounts with low health scores but renewals in the next 90 days, and summarize their recent support ticket themes." 3
  • Synthesize data from multiple sources.
  • Example Prompt: "For Customer Y, show me their current ARR, renewal date, and a summary of the last three Timeline entries." 7
  • Draft detailed, data-rich content.
  • Example Prompt: "Draft an EBR prep email to my internal account team summarizing the key risks, recent successes, and strategic opportunities for this account." 1
  • Get expert CS guidance.
  • Example Prompt: "What are the best practices for re-engaging a customer with low product adoption?" 1

 

Use the Gainsight AI Agent in Slack when you need to...

  • Get a quick fact about a single customer.
  • Example Prompt: "@Gainsight what is the ARR and renewal date for Customer A?" 1
  • Instantly share data with your team.
  • Example Prompt: (In a shared channel) "@Gainsight summarize the last 3 Timeline entries for Customer Z." Then share the result with the team.5
  • Answer a quick request from a colleague.
  • A Sales rep asks for an account's health score. Use the agent in your DM to get the answer in seconds.1
  • Handle an urgent "fire drill."
  • A Product Manager asks about a customer's recent activity. Use the agent to get immediate context without leaving the conversation.5
  • Draft a quick follow-up while in a Slack conversation.
  • Example Prompt: "@Gainsight draft a quick follow-up email to the team about our sync on Customer B's renewal." 5

 

Sources


Thanks for this response, Dayn! When I tried the first prompt you proposed into Cockpit - “Summarize all value delivered for Customer X in the last 6 months, referencing completed CTAs and positive sentiment from Timeline.” - I received this message... “COCKPIT queries are not supported currently”


Hi Roxanne, 

On our end we have launched an internal team of CSMs who are particularly interested in Gainsight and AI. We have tasked them with creating a copilot prompt and best practice resource which they consolidate using their own trial and error or feedback from other CSMs. 

Additionally, we continue to reinforce the importance of logging timeline activities as this effectively allows us to build our own dataset which Copilot will query and sets us up to benefit from future enhancements. 

Hope that helps!


  1. Creating a copilot prompt and best practice resource... 
  1. Continue to reinforce the importance of logging timeline activities… this builds our own dataset which Copilot will query and sets us for success. 

Brilliant.

Recording successful prompts so you’ve got a prompt template library (ex: Google Sheet), and following the “If it’s not in Gainsight (Timeline) it didn’t happen” rule are critical for success.


We are working on a prompt library wherein end users (CSMs ) would be able to save the most used prompts, admins will be able to add prompts for send users to use and there will be couple of prompts from Gainsight end as well which admins can decide to show or not show.