Skip to main content

AI Agents

Scale Stage | $5-15M ARR | 30-80 headcount

Main challenge: Adding capacity without chaos. Process debt and tool sprawl.

← Back to Scale Overview

AI Agent Recommendations

Stage-appropriate approach: AI agents at Scale handle volume and consistency challenges. Automated agents keep processes running at scale. Work agents help leaders and specialists with analysis and strategy.

Automated Agents (Headless)

Use CaseTriggerOutputSOP Required
Lead routingNew lead in CRMAssigned to right rep/queueLead routing rules
Data enrichmentNew contact/accountEnriched recordEnrichment SOP
Activity loggingMeeting/call completedActivity in CRMLogging standards
Report generationScheduled (daily/weekly)Dashboards updatedReport specs
Health score updatesUsage data changesHealth score recalculatedHealth model
Renewal alertsDays to renewal thresholdAlert to CSM/managerRenewal process
Pipeline hygieneStale opportunity detectionAlert or auto-archivePipeline rules

Work Agents (Interactive)

Use CaseWho UsesExample Prompt
Forecast analysisSales leader"Analyze our current pipeline and identify deals likely to slip. Flag where MEDDIC is incomplete."
Deal strategyAE/Manager"Review this enterprise opportunity. What risks do you see? What should I do differently?"
Territory modelingRevOps"Model three territory scenarios based on this account list and rep capacity."
Campaign analysisMarketing"Analyze performance of our webinar series. What's working? What should we stop?"
Churn analysisCS leader"Analyze our churned accounts from Q4. What patterns do you see?"
Competitive positioningSales/PMM"Based on these competitor mentions in calls, what positioning adjustments should we make?"
Process designRevOps"Draft a stage gate process for our sales stages. Include entry/exit criteria."

What's New at Scale (vs. Stabilize)

CapabilityWhy It Emerges
Forecast agentsPipeline volume makes forecasting meaningful
Territory agentsMultiple reps need balanced territories
Health scoring automationCustomer count requires automated scoring
Reporting automationReport volume requires automation
Competitive intel agentsEnough call volume for pattern detection

Implementation Guidance

  • Start with high-volume tasks — enrichment, routing, logging save the most time
  • Build SOPs first — agents execute SOPs, can't work without them
  • Monitor quality — automated doesn't mean unmonitored
  • Iterate — agents improve with feedback and refinement

What NOT to do:

  • AI for AI's sake — automation should solve real problems
  • Skip the SOP — agents without SOPs produce garbage
  • No human oversight — especially on customer-facing outputs
  • Over-reliance early — agents augment, don't replace judgment