CSM call summaries via Gainsight + AI (L2)

The steps

1. 01

   Define the prompt template

   To ensure AI summaries are consistent and actionable, the CS Ops Manager must define a standardized prompt template within the AI tool used for call recording (e.g., Gong, Chorus, or Gainsight’s internal AI). • Navigate to your call recording platform's 'Summarization Settings' or 'AI Prompt Library'. • Create a custom template named 'CSM Executive Summary'. • Use a prompt structured like this: 'Analyze this customer call and extract: 1) Three primary risks (churn indicators or technical blockers), 2) Two expansion or upsell opportunities, 3) A numbered list of stakeholder action items with owners, and 4) Current sentiment sentiment on a scale of 1-5.' • Ensure the output format is compatible with Gainsight 'Timeline' activities (avoiding complex HTML that might break when synced). • Owner: CS Ops Manager. • Time: 2 hours. • Common Pitfall: Being too vague with the prompt, which results in generic 'They had a good chat' summaries. • Definition of Done: A saved prompt that generates 4 distinct sections for every processed call record.

2. 02

   Configure automation sync

   Once the summary is generated, it must be programmatically pushed into Gainsight to save the CSM from manual data entry. • Go to Gainsight Administration > Integrations > [Your Video Tool] or use the Gainsight Horizon Rules Engine. • Map the 'AI Summary' field from your source tool to the 'Gainsight Timeline' Activity Type 'Update' or 'Client Meeting'. • Specifically, map the 'Action Items' section to a custom text field or directly into the 'Internal Notes' section of the Timeline record. • Configure the sync frequency to 'Immediate' or 'Hourly' to ensure the CSM sees the summary shortly after the call ends. • If using Gainsight’s built-in AI, enable the 'Auto-generate Summary' toggle within the Timeline configuration menu in the Administration panel. • Owner: Gainsight Administrator. • Time: 3-4 hours. • Common Pitfall: Forgetting to map the 'Call Date' properly, causing summaries to cluster on the date of the sync rather than the date of the meeting. • Definition of Done: A call summary automatically appears in the Gainsight Timeline for a specific account after a test recording ends.

3. 03

   Automate risk-based CTAs

   To move from passive reporting to active retention management, use the Gainsight Rules Engine to trigger Calls to Action (CTAs) based on keywords in the AI-generated summary. • Navigate to Administration > Rules Engine > Create New Rule. • Set the Data Source to 'Timeline' and filter by 'Activity Type = Client Meeting'. • Add a filter for the 'Comments' or 'AI Summary' text field where it contains high-risk keywords like 'cancel', 'budget', 'competitor', or 'unhappy'. • In the 'Setup Action' section, choose 'Create CTA'. • Configure the CTA: Name it 'Risk Detected via AI Summary', set the Priority to 'High', and assign it to the 'Relationship CSM'. • Include a link to the original meeting recording in the CTA description for context. • Owner: Gainsight Administrator. • Time: 3 hours. • Common Pitfall: Failing to exclude common words that might trigger false positives, leading to 'CTA fatigue' for the CSMs. • Definition of Done: A risk-specific CTA is automatically created in a CSM's Cockpit when a test summary contains the word 'cancelled'.

4. 04

   Conduct manager quality audit

   CS Managers must perform a weekly quality audit to ensure the AI and the CSMs are acting on the insights provided. • Every Friday, the CS Manager should open the 'Gainsight Cockpit' or a custom 'Activity Report'. • Use a filtered view to see all 'AI-Generated Summaries' from the past 7 days. • Randomly select 3 summaries per CSM. • Review the summary against the recorded call (watch at 2x speed) to check for accuracy. • Check if the CSM updated the 'Company Health Score' or 'Success Plan' based on the findings in the summary. • Provide feedback in the 'Timeline' comments section of the record to show the CSM that the leadership team is monitoring the data. • Owner: CS Manager. • Time: 1 hour per week. • Common Pitfall: Checking only the summary and not verifying if action was actually taken on the tasks listed. • Definition of Done: Three annotated reviews completed per CSM in Gainsight, with documented feedback on summary accuracy.

5. 05

   Measure admin time and quality

   The final step is to measure the impact of the AI on the team’s efficiency and the quality of their account management. • Create a Gainsight Report using the 'Timeline' and 'CTA' objects. • Metric A (Efficiency): Average time spent on 'Admin' tasks per CSM (survey the team monthly to compare against the pre-AI benchmark). • Metric B (Actionability): Percentage of AI-flagged risks that were closed within 14 days versus manually flagged risks. • Metric C (Coverage): % of total customer meetings that have a completed AI summary in the Timeline. • Present these findings in the monthly RevOps/CS leadership meeting to justify the continued use of AI seats. • Owner: RevOps Analyst. • Time: 2 hours (setup), 30 mins (monthly review). • Common Pitfall: Focusing only on 'CSM time saved' without checking if 'CTA Quality' is improving retention outcomes. • Definition of Done: A live Gainsight Dashboard showing the 'AI Summary Adoption' and 'Risk Resolution' rates.

Next playbooks

• AI-generated QBR decks (L3)
• Intercom Fin AI for support deflection (L3)
• Support deflection with AI search (L3)
• Customer health scoring with AI signals (L4)