User Segment

User segments exist only as mental models in people's heads. The head of marketing says 'our power users' but cannot articulate exactly which criteria define that group. When sales asks 'who are our enterprise-ready accounts?' the answer is intuitive, not documented. There are no written user segment definitions anywhere in the organization.

None — AI cannot target or analyze user cohorts because no user segment definitions exist in any system or document.

User segments are defined in scattered places — a marketing deck lists 'personas', the analytics tool has saved filters, and sales has an ICP spreadsheet. Definitions conflict: marketing's 'enterprise' user segment means 500+ employees while sales defines it as $100K+ ARR potential. Nobody has reconciled these competing user segment definitions, and each team operates with its own version of the truth.

AI could surface the various user segment definitions, but cannot reliably apply them because the same segment name maps to different criteria depending on which team's definition is used.

User segments have canonical definitions in the analytics platform with documented criteria — behavioral thresholds, firmographic filters, and engagement scores. Teams reference the same user segment definitions. But the definitions are static descriptions without links to the underlying metric calculations or behavioral event sources. When someone asks 'what exactly counts as an active user in this segment?' tracing the definition back to raw events requires manual investigation.

AI can apply user segment definitions to filter users and generate cohort reports. Cannot validate whether a segment's criteria still reflect the intended business concept because definitions lack traceability to the underlying event schemas and metric formulas.

User segments are fully documented with traceable criteria. Each user segment record links to the specific behavioral events, metric calculations, and threshold values that define membership. A product analyst can query 'show me the exact event sequence and metric thresholds that define our At-Risk user segment' and get a complete, auditable answer. Segment definitions are current and version-controlled.

AI can audit user segment definitions for logical consistency, detect when underlying metrics shift enough to change segment composition materially, and recommend criteria adjustments. Cannot yet auto-optimize segment boundaries because definitions lack formal relationships to the outcomes they predict.

User segments are formal entities in a customer intelligence ontology. Each segment has typed relationships to defining criteria, predicted outcomes, overlapping segments, and the campaigns or product experiences that target it. Segment criteria are machine-readable specifications. An AI agent can answer 'which user segments predict churn with more than 80% accuracy and are not currently targeted by a retention campaign?'

AI can autonomously evaluate user segment effectiveness against outcome metrics, propose new segment definitions based on behavioral clustering, and detect when existing segments have drifted from their original predictive power.

User segments are self-documenting and self-calibrating. Segment definitions continuously update their criteria based on observed outcomes — the 'At-Risk' user segment automatically adjusts its behavioral thresholds as churn patterns evolve. New segments emerge automatically from behavioral clustering, complete with generated names, descriptions, and predicted outcome associations. The user segment catalog is a living taxonomy that reflects current reality without manual definition updates.

Can autonomously discover, define, calibrate, and retire user segments based on behavioral patterns and outcome correlations. AI maintains the complete segment taxonomy as a self-evolving system that always reflects current user behavior.

Ceiling of the CMC framework for this dimension.