AI Monetization Is a Continuous Experiment. Here’s How to Run It Well.

What does it look like to price AI in a way that satisfies your customers and helps you scale? That question is driving some of the most important conversations in technology right now, and Chargebee’s own research points to a clear answer: the companies building pricing in lockstep with their products are growing twice as fast as those that don’t.

In the first session of Chargebee’s Monetization Gamechangers series, Lily Varon, Principal Analyst at Forrester,  joined Chargebee CMO, Guy Marion, to work through exactly that. Here are some of the top takeaways to share with your team.

Seats, Usage, Outcomes: AI Pricing Models Are a Menu, Not a Linear Path

The prevailing wisdom has it that seat-based pricing is dying, usage-based is the bridge, and outcome-based is the inevitable destination. Lily pushed back on that. “It’s a menu,” she said. “Subscriptions, usage, and outcomes are all going to coexist. The right model depends on your product, your customer, and your ability to attribute value.”

Salesforce’s AgentForce journey is worth paying attention to. The product launched at $2 per conversation, with a clean usage model. Salesforce then moved toward outcome-based pricing. By December 2025, Mark Benioff signaled on an earnings call a potential return toward seat-based pricing, driven by what customers were asking for: predictability. That’s a company doing exactly what the moment requires: experimenting, listening, and adjusting. There is no final destination. There’s only the discipline to keep iterating.

Chargebee’s State of Recurring Revenue & Monetization report, drawn from a survey of over 450 North America and EMEA-based subscription and recurring revenue businesses, found that companies adding AI to their products that keep pricing strategy aligned with product development are twice as likely to be growing fast as those that treat pricing as a separate conversation. Just as product teams never stop improving their products, monetization never stops either.

Why Outcome-Based AI Pricing Is Harder to Execute Than It Sounds

Outcome-based pricing has obvious appeal: charge for value delivered, not access granted. It aligns incentives, builds trust, and removes the ceiling on what customers will pay as AI gets more capable.

The execution is where things get hard.

Guy Marion knows this firsthand. At Brightback, the churn prevention company he founded and later brought to Chargebee as Chargebee Retention, the team built a pricing model tied to revenue saved from reduced churn. Customers pushed back almost immediately. Would that subscriber have churned anyway? Would they have come back on their own? The attribution question stops being theoretical very quickly.

Lily pointed to professional services as a useful reference. Consultancies have been running outcome-based experiments for years, well before AI entered the vocabulary. What they found consistently: buyers and sellers end up arguing about what the outcome was, whether it was achieved, and who deserves credit for it.

That said, Lily identified four signals that suggest outcome-based pricing may be a good fit for your business:

These aren’t guarantees. They’re starting points. Check all four, and the conversation is worth having seriously. Miss one, and calibrate your expectations that it may not be worth moving to outcome-based pricing

Your Pricing Strategy Needs to Move as Fast as Your Product Does (Maybe Even Faster)

Pricing can’t be a launch decision that gets revisited once a year. AI infrastructure costs can drop 40% in a matter of months. Customer willingness to pay is still being established. Your pricing has to move at the same pace as your product.

The State of Recurring Revenue & Monetization report found that 80% of software companies adding AI to their products are evolving their pricing in regular lockstep with product development. In practice, that means:

Guy cited an AI video company whose per-video production costs for a seven-second video dropped from $5 to under $3 in under six months due to sinking LLM costs. They were able to adjust their pricing quickly in response because their monetization infrastructure was built for change, not locked into it.

CFOs Now Have a Seat at the AI Pricing Table. Many Aren’t Sitting At It Yet.

As agentic AI actions become harder to forecast and AI infrastructure costs remain unpredictable, CFOs are stepping into a more active role in pricing decisions. That shift is overdue.

Your billing and monetization stack should be treated as a strategic asset. It’s where usage data is captured, where rate logic is applied, where revenue is recognized, and where customer disputes either resolve cleanly or turn into relationship problems. None of that is back-office work anymore. It requires tight alignment among finance, product, and go-to-market, and a CFO who understands not just the numbers but the commercial model generating them.

This doesn’t mean every CFO needs to become a technologist. It means the CFO needs to be in the room when pricing decisions happen (cue Hamilton music), not just when the results come in.

The Companies Getting AI Monetization Right Share a Common Approach

There is no single right answer to AI monetization. What matters is building the discipline to experiment with pricing, test with customers you trust, and stay flexible enough to change course when the market moves.

Lily put it simply: go back to first principles. Know what value you deliver, find the unit that proves it, and build your commercial model from there. The hype will keep moving. The fundamentals won’t.

Guy and Lily go deeper on all of this in the full session, including specific frameworks for evaluating outcome-based pricing fit and what your tech stack needs to support rapid monetization iteration.

On-Demand Webinar

Watch the full Monetization Gamechangers session