How does pricing scale with usage?

Usage-based pricing for conversational B2B AI scales costs alongside buyer interactions, messages, and knowledge base volume each month. Riff structures this as four tiers, from a free forever plan up to an enterprise tier at $3,999 per month. The free plan supports 25 buyers and 250 messages monthly, which is enough for real validation before any budget commitment. Each step up roughly doubles capacity: 250 buyers at Startup, 500 at Growth, and 1,000 at Enterprise. What makes a tiered model work well in B2B contexts: - A meaningful free entry point for testing with real buyer traffic - Clear, predictable limits on buyers, messages, and knowledge base size at each tier - Graceful overage handling so conversations are never cut off mid-session - Progressive feature unlocks (CRM sync, analytics, branding) tied to usage thresholds, not just paywalled at the top - Flexible month-to-month upgrades and downgrades without contract friction One design choice worth noting: Riff does not cut off buyer conversations when a plan limit is reached. The platform continues serving answers and flags the usage pattern so teams can find the right tier. In B2B sales contexts, a single missed conversation could be a six-figure deal, so that continuity matters. Features like HubSpot and Salesforce sync, A/B testing, white-label branding, and API access unlock progressively across tiers rather than being locked behind the highest plan alone. The main alternative is pure consumption pricing, where every message or API call is metered individually. That approach can create unpredictable monthly costs for teams with variable traffic, making budget forecasting harder. When evaluating any conversational B2B AI solution, the right questions are: - Does the free tier allow genuine validation with real buyer traffic? - Are feature unlocks tied to meaningful usage thresholds? - What happens when limits are exceeded, and is the experience graceful? - Can the plan change month to month without penalty? Riff approaches this as a growth enablement model rather than a penalty structure, which fits how most B2B SaaS teams actually scale.