REACH Scoring Agent
A Glean Agent that scores accounts across 5 dimensions on a 90-day cadence and key events — with human review, one-click Totango import, and historical score tracking.
The Problem
Account prioritization depended on gut feel
Subjective judgment and inconsistent criteria made it impossible to scale CSM execution across a large portfolio.
Across a $250M ARR organization with 1,200+ accounts, Customer Success Managers relied on gut feel and inconsistent criteria to prioritize their time. Without a standardized scoring model, high-value expansion opportunities were missed and at-risk accounts slipped through the cracks. Manual scoring was too slow to be useful — by the time a CSM finished their review, the signals had changed.
How It Works
Scheduled scoring, human confirmation
Scheduled & event-triggered runs
The agent runs automatically on a 90-day cadence for each account, and is also triggered by key events: objectives achieved, business reviews completed, or new scores received. CSMs can also run it on-demand for any account at any time.
Slack notification with linked review
When a scheduled run completes, the CSM receives a Slack DM confirming the results are ready. The message includes a direct link to the generated Glean conversation where the full scorecard, dimension breakdowns, and justifications are laid out for review and editing.
One-click Totango import
After reviewing and optionally editing the scores and justifications, the CSM confirms with a single button click. The results are imported into Totango as live scores against each REACH dimension — with historical time tracking so every score change is logged and visible on a timeline graph.
The Framework
REACH: 5 dimensions, deterministic scoring
Adapted Rod Cherkas's REACH™ framework — designed for post-sale expansion revenue — into a quantifiable scoring engine. Each dimension is mapped to data signals from Salesforce, Totango, Provisio, Gong, GSuite, Slack, and Qualtrics (NPS via SFDC), producing a scorecard CSMs can act on immediately.
Relationships — Stakeholder Mapping
30%Mapped the depth and breadth of customer relationships — number of engaged stakeholders, executive sponsor connections, and cross-functional touchpoints. Scored stakeholder coverage against account tier, flagging single-threaded relationships at highest renewal risk. CSMs could sort by R-score to prioritize accounts with thin executive coverage before contract renewal.
Engagement — Beyond Usage Metrics
20%Measured engagement as a leading indicator of expansion — not just login counts, but meaningful interaction patterns: feature exploration, training participation, Gong call cadence, and proactive support outreach. Accounts with rising E-scores were flagged for upsell conversations; declining scores triggered re-engagement plays.
Actions — Signal Capture
25%Trained the scoring model to detect small expansion signals before they became obvious — feature requests in support tickets, org changes in Salesforce, budget cycle timing, and champion tenure shifts. The system captured these actions and prompted CSMs to act within 48 hours, turning weak signals into expansion opportunities.
Customer Value — Executive Resonance
15%Scored how effectively CSMs demonstrated measurable ROI to executive stakeholders — QBR completion rate, documented business outcomes, NPS/CSAT trajectory, and executive meeting frequency. Accounts with low value resonance were flagged for executive sponsorship intervention before the relationship deteriorated.
Horizons — Future Growth Mapping
10%Assessed expansion opportunity across new business units, geographic regions, and product lines. Scored white-space ARR, product tier ceiling, upsell signal recency from Gong calls, and customer headcount growth. High H-scores surfaced accounts where AE-led expansion conversations were likely to succeed.
Visual Architecture
REACH: 5 dimensions, deterministic scoring
Scoring Model
Deterministic logic, not black-box AI
Each REACH dimension uses a base score plus additive modifiers tied to concrete data signals — not opaque ML weights. This makes every score auditable and explainable to CSMs and leadership. The composite REACH score is a weighted average (R×0.30 + E×0.20 + A×0.25 + C×0.15 + H×0.10), with segment classification (GREEN ≥ 75, YELLOW ≥ 50, ORANGE ≥ 30, RED < 30) driving automated CSM plays.
Results
Execution focus at scale
1,200+
Accounts scored on cadence
$250M
ARR org-wide impact
One
Click to import into Totango
Connected Systems
Agent Framework
Methodology Phases