WORKFLOWSeparate application or agent-native skill?
A conventional platform centralizes administration and reporting. An agent-native skill reduces context switching and makes natural-language tasks portable. Neither architecture is universally superior; evaluate governance and operator behavior.
DATASingle source or waterfall enrichment?
A single source can simplify provenance and predictable cost. Waterfall enrichment may expand coverage while introducing provider conflicts and variable cost. Ask how every field identifies its origin.
CONTROLAutonomous sending or human review?
Autonomy can increase throughput. Review gates protect relevance, regional policy, sender reputation, and brand voice. This implementation assumes people approve targeting, messaging, and sending.
COSTSeats, credits, providers, and runtime
Total cost can include user seats, data credits, verification, APIs, infrastructure, and operator time. Compare a representative task rather than a headline price; mark unavailable inputs Unknown.