How do I choose a pricing model for my AI product?

Choosing Your Billing Unit

One of the biggest decisions when monetizing an AI product is: what do you charge for? This question comes up in almost every customer conversation we have. Here are the most common approaches and when each makes sense.

Option 1: Per-Request Pricing

Charge a fixed amount per API call, investigation, or action.

Best for: Products where each action has roughly equal cost (chatbots, simple agent calls)
Pros: Simple for customers to understand. "Each investigation costs X credits."
Cons: If some requests are much more expensive than others (e.g., complex multi-step agent calls vs. simple lookups), your margins become unpredictable.

Option 2: Token-Based Pricing

Charge based on actual AI token consumption with a markup.

Best for: Products with highly variable request complexity
Pros: Costs directly reflect actual AI usage. Your margins stay consistent regardless of request complexity.
Cons: Customers may not intuitively understand token counts. Can feel unpredictable if usage varies a lot.

Option 3: Credit-Based Pricing

Sell credits that map to your own unit of value, with different actions consuming different amounts of credits.

Best for: Products that want to abstract away the underlying AI costs
Pros: Full control over pricing. You can adjust credit consumption rates for different features without changing customer-facing prices.
Cons: Customers sometimes feel confused when credits deplete faster than expected if the mapping isn't transparent.

Option 4: Subscription + Overages

Charge a monthly subscription that includes a set amount of usage, with overage charges beyond that.

Best for: Products transitioning from seat-based SaaS pricing to usage-based
Pros: Predictable base revenue. Familiar to customers used to SaaS subscriptions.
Cons: You need to set the included usage amount carefully. Too generous and you lose margin; too stingy and customers churn.

Lava Supports All of These

With Lava, you don't have to commit to one model upfront. You can:

Configure meters for any billing unit (requests, tokens, custom credits)
Set fixed or percentage-based markups
Add tiered pricing with volume discounts
Combine subscriptions with usage overages
Change pricing in the dashboard — no code changes needed

This flexibility lets you experiment quickly. Start with one model, measure how customers respond, and adjust without rewriting billing code.

Our Recommendation

If you're not sure, start with subscription + overages. It gives you predictable revenue while letting customers who use more pay more. As you learn your customers' usage patterns, you can fine-tune the included amounts and overage rates, or switch to pure usage-based pricing later.