Pricing and Packaging
Design AI-era SaaS pricing models, packaging, margins, and governance.
Key takeaways
- AI changes pricing: seat-based pricing alone fails to capture AI usage and value, pure usage pricing makes budgets unpredictable, so most B2B AI SaaS needs a hybrid model.
- Price across three axes: access (who and what features), consumption (how much), and value (what work outcome).
- Pick a customer-understandable value metric instead of passing token or execution cost straight through.
- Check gross margin after AI execution and service cost before publishing a price.
- Use AI for discount approval, package recommendation, and renewal pricing, and validate packaging changes by cohort adoption and churn.
AI changes pricing strategy. Seat-based pricing alone often fails to explain AI usage, cost, automation value, and outcomes. Pure usage pricing can make budgets unpredictable. Most B2B AI SaaS needs a hybrid model.
Pricing Model Comparison
| Model | Strength | Risk | Best fit |
|---|---|---|---|
| Seat-based | Predictable and easy to buy | Weak link to AI usage or value | Collaboration and admin-heavy products |
| Usage-based | Connects cost and value | Harder budget forecasting | API, data processing, automation |
| Credit-based | Packaging flexibility | Can feel complex | Bundled AI features |
| Outcome-based | Direct value link | Attribution disputes | Clear cost reduction or revenue uplift |
| Hybrid | Balances predictability and scale | Harder to explain | Most B2B AI SaaS |
Three Pricing Axes
- Access: who can use which features
- Consumption: how much is used
- Value: what work outcome is achieved
Good packaging explains all three without making budget planning impossible.
Value Metric Selection
| Product type | Possible value metric |
|---|---|
| Sales AI | Accounts, prospected leads, meetings assisted |
| CS AI | Active customers monitored, tickets resolved, playbooks run |
| Data AI | Rows processed, workflows, data sources |
| Document AI | Documents analyzed, seats, projects |
| Agent platform | Tasks, runs, credits, connected systems |
Cost pass-through is not enough
Passing token or execution cost directly to the customer makes pricing feel supplier-centered. Connect cost units to customer-understandable value units.
AI Package Margin
AI features carry token, retrieval, tool-call, and human-in-the-loop costs. Before publishing a price, check gross margin after AI execution cost and service cost.
Manage Discounts and Exceptions With AI
| Area | AI role |
|---|---|
| Discount approval | Compare similar deals, margin, and win probability |
| Package recommendation | Use usage patterns and expansion potential |
| Renewal pricing | Reflect adoption, value proof, and risk |
| Pricing experiments | Detect segment response and churn |
| CPQ support | Generate guidance and approval rationale |
Packaging Principles
- Let buyers experience core value in free or lower tiers.
- Put expansion value in teams, data, automation, security, and governance.
- Use fair-use and overage rules for expensive AI features.
- Make enterprise tier security, audit, admin, and SLA explicit.
- Price pages should explain the operating problem solved, not only included features.
Operating Checklist
- Pricing reflects both AI cost and customer value metric.
- Customers can forecast monthly cost.
- Discount approval includes margin and value proof.
- Packaging changes are validated by cohort adoption and churn.
- Pricing experiment results update sales enablement.