What’s monetized is changing in the AI world. No longer is software monetization based strictly on access and users, but increasingly on AI transactions, duration of usage, specific feature consumption, API calls, data processing, and more.
Today’s software producers are grappling with the financial impact of variable AI infrastructure costs, unpredictable usage patterns, and evolving customer expectations—all of which are putting pressure on traditional pricing models.
Usage-based pricing is emerging as a natural fit for AI, helping producers recover costs while aligning price with value. Many are layering consumption components onto core subscriptions—anchoring predictable revenue while monetizing AI capabilities as usage scales.
When designed around clear value metrics and supported by the right operational foundation, this hybrid approach creates more durable revenue while delivering flexibility and transparency to customers. Business transformation is essential to support this shift.
Here’s my 6-step framework to evolve your pricing, systems, and processes to make AI monetization successful.
1. Prioritize customers’ needs.
Customer considerations always need to be at the top of the list. Customers must be able to budget. They need transparency about their entitlements and what they’ve used.
Predictability is super important. Don’t introduce a new monetization model without providing visibility into usage history. Customers won’t accept a consumption model if they don’t have the opportunity to see, in real time, how much they’ve already used—particularly when it comes to high-cost features, such as AI workloads. They’ll want end-user controls, the ability to define who (or which department) can (or can’t) use credits or tokens. If customers can’t predict expenses, expect pushback.
For current customers, measuring consumption for about 6 months—and communicating it clearly with usage reporting—provides the insights needed to support buy-in, essentially warming them up before transitioning to the new usage-based pricing strategy.
2. Define “usage”!
What is the usage that you’re going to count? How are you going to count it?
These seemingly simple questions prove to be quite challenging for businesses that are shifting to usage-based monetization models. To avoid contention, disputes, and misunderstandings, stakeholders must agree on what “usage” or “consumption” actually is.
Businesses must evaluate variables including:
- The model: Is your monetization strategy prepaid or postpaid? Is the amount of usage included with a subscription, but capped at a certain amount? Is usage offered in exact amounts or in tiers that provide particular volumes (such as a good/better/best model)? What compliance considerations must be addressed, such as how overages and/or grace periods will be defined and enforced (and the potential impact on business relationships)?
- The metric: Is the metric being monetized based on data usage, per-transaction, access, or outcomes? Will this work for most customers or are usage scenarios so diverse that additional adaptations are necessary?
- The term: Does the consumption that’s being paid for expire or not? For how long can it be used? How do you implement the model when selling through a channel?
3. Get executive buy-in.
Rest assured: different people will have different opinions. Without clarity and executive alignment, the launch of usage-based pricing for AI capabilities may very well be delayed. Make sure there’s agreement on what is getting monetized and how it’s done.
- The CRO or sales leader needs to retrain the salesforce on how to position and sell consumption-based components, and how to work with customers in such a context.
- The CFO needs to make sure that the company grows and that revenue is predictable. Internal work is essential to take into consideration the different priorities of different departments—and to deliver what customers want, need, and value.
- The CPO wants to drive adoption, so needs clear insight into how products are used and which capabilities are leveraged—even when consumption credits can be applied flexibly across a range of applications.
Usage-based pricing impacts sales methodology, annual recurring revenue (ARR), customer churn, revenue recognition, and financial/accounting processes. Because of this widespread impact, many efforts to shift to usage-based monetization fail if the C-suite isn’t completely on-board. To guard against this, bring sales and finance teams into the plans early. Model a variety of scenarios, including best-case, medium, and conservative possibilities; be transparent about potential revenue dips and volatility.
4. Use the right technology.
I often see companies rushing to implement new monetization and pricing models—and building something quickly, with whatever system is available. This creates a lot of problems down the road. Too often this is only a bandage, not the full strategy necessary to address the evolving considerations for how to monetize AI.
To avoid quick fixes that ultimately lead to complications and technical debt, consider:
- How well does the solution scale?
- What customer experience does it offer across products?
- How well do the end user controls provide transparency to customers?
- How effectively does it measure usage of AI both online and offline (such as on devices that aren’t permanently connected to the internet)?
- Does it provide the reliability, stability, and audit trails your organization needs?
5. Make decisions based on data, not assumptions.
As you prepare for a consumption-based model of your AI-based products and features, you need to be data-driven. Making decisions based on what product managers assume, rather than what customers actually use, is an unnecessary risk.
Not everything lends itself to a consumption model. To see if what you’re monetizing does, dive into telemetry. Track consumption; evaluate usage patterns (such as those based on seasonality, industry, customer type, and/or customer size). Have clear customer and product mapping. Evaluate different timeframes and customer cohorts. Analyzing and understanding this data is crucial to forming a profitable AI pricing strategy.
6. Start small.
Launching with a big bang likely isn’t practical. Roll out with a small set of customers or with a minimum viable product (MVP). This allows you to test assumptions, see how well they work, and grow from there. Apply a simple rate table that breaks down the consumption of different AI features and what they cost; these can be adjusted as needed.
—
Once you launch a consumption-based model, it will continue to change. Build into your project planning a cycle in which you review, iterate, and refine on an ongoing basis in order to support the needs of your customers, as well as the needs of your finance and sales teams. Stay focused on the convergence of cost and value—and you’re likely to be successful.
