Configuration Best Practices
This article outlines configuration best practices for using Staircase AI in Gainsight CS and standalone Staircase AI instances to enhance customer intelligence, automate risk detection, and drive proactive account engagement.
Overview
Staircase AI transforms everyday customer interactions into actionable intelligence that helps you drive growth, reduce churn, and strengthen customer relationships. By analyzing sentiment, engagement, and relationship dynamics across communication channels, Staircase AI equips your team with real-time insights to act with confidence and precision.
This best practices guide outlines how to effectively use Staircase AI integrated directly with Gainsight CS, to support proactive workflows that turn insights into outcomes, automatically and at scale.
What You Can Achieve with Staircase AI
- Track Customer Signals Automatically: Monitor customer health, sentiment, and engagement continuously with no manual tracking required.
- Identify Expansion Opportunities Early: Detect upsell and cross-sell signals embedded in customer conversations and surface them in your workflows.
- Understand Relationship Dynamics: Gain visibility into stakeholder alignment, engagement strength, and influence on account retention or growth.
- Take Immediate, Informed Action: Use built-in integrations to trigger CTAs, alerts, and playbooks based on AI-driven insights, ensuring the right follow-up at the right time.
- Earlier Risk Detection: Detects sentiment shifts, competitive mentions, and relationship changes, up to six weeks earlier than usage data alone.
Recommended Configuration Flow
To use Staircase AI effectively, complete the following steps in order:
- Set Up Your Organization Domain
- Review Privacy and Security
- Configure Tier-Based Health Scores
- Configure Team Effort
- Set Up Stakeholder Roles and Engagement Cadence
- Configure Tier-based Insights
- Map Journey Phase
- Configure Insights
- Configure Lifecycle Events
- Configure Topics
- Set Up Smart DM and Channel Alerts for Account Monitoring
Set Up Your Organization Domain
Read and validate the profile summary and if needed update it if anything is incorrect. Before onboarding teams and loading any data to Staircase AI set up your organization domain:
- Organization Domains: Add internal email domains, for example, `@company. com`under Add organization domains. Define which domains should be treated as part of your organization. Emails from these domains are recognized as internal communications and internal communications are not pulled in and analyzed by Staircase AI.
- Domain Exclusion List: Add domains that should never be analyzed, such as legal, HR, or partner domains. Communications from these domains are excluded even if a customer domain is included.
- Trusted Login Domains: Add other domains that can access the Staircase AI in Trusted Login Domains. These users must be invited or synced through SSO. Communications from these domains are not tracked as internal activity, unless they are added to the internal domain list.
Audit domains quarterly, especially after acquisition activity or changes in partnerships.
For more information, refer to the Organization article.

Review Privacy and Security
Use the ‘Are email bodies visible to other users?’ setting to control transparency:
- Enabled (Recommended): All users can view email content regardless of role or if they are on the communication thread.
- Not Enabled: Visibility is restricted based on user roles and participation.
Choose the option that best aligns with your internal privacy policies. For more information, refer to the Privacy & Security article.

Configure Tier-Based Health Scores
Calibrate your health scoring logic by customer segment to reflect how different customer types engage. Shorter windows for high-touch accounts and longer ones for scaled segments help reduce false positives and ensure signals stay accurate and actionable. Create a separate health score for each tier—Enterprise, Mid-Market, and SMB/Digital—using appropriate lookback windows and weight distributions.
Recommended Lookback Windows:
- Enterprise: 30 days
- Mid-Market: 60 days
- SMB/Digital: 90 days

Shorten windows for high-touch accounts
- These customers interact frequently, so a shorter window captures meaningful changes quickly.
- Use the OOTB weighting for all tiers except SMB/Low touch.
Balance signal and noise in the mid-market
- Mid-Market: Use a 30–60 day window.
- Keep the same weighting mix but monitor whether engagement is truly representative, example, sentiment (35%), engagement (40%), response time (25%).
Recommendation: Keep the out-of-the-box (OOTB) weighting for all segments except SMB. For SMB accounts, reduce the weighting on engagement and redistribute it across the other three inputs. This assumes sentiment analysis for support tickets is enabled. If sentiment is not enabled, reduce weighting on both engagement and sentiment, and redistribute that weight across open items and response time. Consider layering in a product usage signal for additional context. - This tier typically receives more frequent engagement, so a shorter window offers more accurate, real-time insights.
Smooth out variability for low-touch accounts
- SMB / Scaled: Use a 90 day lookback.
- Reduce engagement weighting to ~25–30% and reallocate the remaining 10–15% across other metrics such as sentiment, response time, and open items.
- Sparse interaction means short windows create misleading spikes.
Use health scores as designed; if adding more metrics, limit to two to three and ensure they are numeric. For example: product usage signals, not NPS or CSAT.
Best Practice: Create a product score in your usage analytics tool and pull that product health score into Staircase AI. Ideally, display the product score alongside the Staircase AI health score for a more complete view. For SMB or digital-led segments, where human engagement is limited, it’s recommended to increase the weight of the product signal and reduce emphasis on engagement-related metrics. This approach is optional, but Gainsight recommends applying it only to the SMB tier for accurate signal calibration.
We do not recommend changing the default configuration.
For more information, refer to the Staircase AI Health Score article.
Configure Team Effort
Estimate the team’s average turnaround time for tasks in Settings > Configurations > Team effort.

Recommended per-exchange durations for Email, Chat and Tickets is 3 to 25 minutes. Do not use resolution time, this duration is per message/ticket/chat response.
Optionally, If you would like to account for internal preparation time, add two to five minutes to each field.
For Account manager cost, use the average CSM or account manager salary, ideally from a single region, for example, US average for consistency.
For more information, refer to the Team Effort Hours article.
Set Up Stakeholder Roles and Engagement Cadence
Gainsight recommends mapping stakeholder roles from your CRM. Assign engagement frequency for key roles to trigger insights and CTAs:
| Role | Recommended Touch Frequency |
|---|---|
| Champion | Weekly / Biweekly |
| Decision Maker | Monthly |
| Executive Sponsor | Quarterly |

If a role does not have an assigned cadence, insights related to that role do not trigger.
For more information on how to assign touch frequency, refer to Stakeholders article.
Configure Tier-based Insights
Customer engagement frequency varies by account segment. To ensure insights remain meaningful and actionable, configure tier-based rules that align with your go-to-market (GTM) model. This approach prevents over-alerting on low-touch accounts and under-flagging risk on strategic ones, enabling your teams to focus on the right accounts at the right time.
Key Benefits
- Reduce alert fatigue: Avoid unnecessary alerts for SMB or low-touch accounts.
- Improve accuracy: Ensure insights trigger appropriately for each account segment.
- Strengthen alignment: Match insights to your organization’s engagement strategy and customer tiering model.
Align Insights to Your GTM Model
Define account tiers such as Enterprise, Mid-Market, and SMB based on CRM fields or Gainsight CS Rules. Consistency in tier definitions ensures insights trigger evenly across teams and eliminates noise caused by uneven configurations.
Tier-based insights help maintain the right level of attention across customer segments. By tailoring insight triggers by tier, Gainsight users can ensure alerts are relevant, actionable, and aligned with their organization’s customer engagement model.
The following table displays tier-based insight differentiation:
| Tier | Engagement Rule | Purpose |
|---|---|---|
| Enterprise (High-Touch) | Flag No Activity after 14 - 30 days | Keeps engagement cadence tight to match frequent customer interactions. |
| Mid-Market | Flag after 30-60 days | Balances meaningful signals with manageable workloads. |
| SMB / Scaled | Flag after 90 days | Prevents excessive alerts for lower-touch accounts. |
| QBR Insights | Enable only for Enterprise and Mid-Market accounts | Quarterly reviews are expected only in these segments. |
For more information on configuring Tier, refer to the Tier article.
Map Journey Phase
Map journey phase from your CRM for consistency and to power relevant insights.
For more information on configuring the journey phase, refer to the Journey Phases article.
Configure Insights
Review each OOTB Insight and adjust based on relevance to tiers:
- Enable or deactivate individual insights.
- Apply filters to restrict insights to specific tiers. For example, enable No QBR only for Enterprise and Mid-Market.
- Configure lookback windows to determine insight thresholds. For example, 365 days for Enterprise QBRs, 180 for SMB.
Configure insights first before enabling them for team visibility to avoid irrelevant CTAs.
For more information on insights, refer to the Insights article.

Configure Lifecycle Events
Enable or deactivate lifecycle events based on triggers to get a complete view of customer engagement. Use Exclude Lookup Terms for false positives. For example, filter out cancel invoice from churn indicators if applicable.
For more information on customizing lifecycle events, refer to the Lifecycle Events article.

Configure Topics
Review and customize out-of-the-box topics based on your requirements. You can enable or deactivate these topics as needed, write clear prompt-based descriptions instead of relying on keywords, and add custom topics using language prompts to further refine topic classification.
For more information, refer to the Topics article.

Set Up Smart DM and Channel Alerts for Account Monitoring
Stay ahead of customer risks and opportunities with smart, tier-based alerts delivered using Slack direct messages (DMs) and dedicated channels. These targeted alerts help account owners, managers, and cross-functional teams take timely action while avoiding alert fatigue.

Direct Message Alerts for Assigned Team Members
Configure DM alerts based on account tier, so each team member receives the right level of signal strength for their portfolio.
- High-Tier Accounts: All key signals (High Noise) including urgent and relationship-critical indicators:
- Churn Risk
- Account Dark (15–30+ days)
- Executive Ghosting
- Negative Sentiment (remove the support channel)
- Stakeholder Changes
- Milestone Delays. For example, no QBR scheduled, no renewal discussion
- Mid-Tier Accounts: Critical signals (Medium Noise) focusing on severe changes:
- Churn Risk
- Account Dark (14+ days)
- Negative Sentiment Trend (remove the support channel)
- Low-Tier Accounts: High confidence signals (Low Noise) delivering only the strongest indicators:
- Churn Risk
- Designed to keep noise low while maintaining early warning coverage.
- Negative Sentiment Trend (support channel included)
Channel Alerts for Non-Assigned Teammates
Smart alerts can also be routed to specific channels for broader visibility across roles.
- Manager Portfolio Channel (Medium Noise)
- Churn Detected, Churn Risk Detected, Extremely Negative Message and so on.
- Executive Channel (Low Noise)
- Revenue-Threatening Scenarios in Strategic Accounts
- Sustained Negative Sentiment Trends
- Support Escalation Channel (Medium–High Noise)
- Churn Risk
- Extreme Negative Sentiment in Support Tickets
- Wins Channel (High Noise)
- Positive Signals: Renewals, Expansions, Customer Praise
- Ideal for boosting morale and team visibility into success moments.
- Set DM rules by account tier using the alert configuration UI. For visibility send as Slack or Teams notification. However for action required by CRM/CSM send the notification as a field to CRM/CSP to initiate follow-up. For example, send Dark Account as a Boolean field to the CRM, and use logic there to create a task.
- Assign team members to receive alerts based on account ownership.
- Use channel-specific filters to avoid duplicating signals across workstreams.
For more information on setting up notifications, refer to the Setting Up Email and Slack Notifications in Staircase and Setting Up Notifications for Microsoft Teams articles.