4 AI Features Monetization Models
How to monetize your AI features as a non-AI-native company?
Hey, Pierre-Jean here 👋 Welcome to this new edition of The Growth Mind!
Today, we discuss:
How AI features are becoming a must-have
The 4 AI features monetization models
A framework to choose the right AI monetization strategy
[Sponsor] Want to benchmark your digital maturity? Take the free Amplitude assessment
Understanding your digital maturity is the first step to improving it. But without the right framework, it’s hard to know where you stand.
That’s why Amplitude built an interactive Digital Maturity Assessment: a quick quiz that evaluates your company’s strengths across 4 key pillars:
Strategic alignment – Are teams working toward a shared vision?
Organizational readiness – Is data at the core of decision-making?
Operational readiness – Can teams move fast with automation?
Technology readiness – Is your stack built for scale and innovation?
In just a few minutes, you’ll get:
✅ A personalized score across the four pillars
✅ Identified gaps and areas to improve
✅ A custom report with industry best practices
AI features are becoming a must-have for non AI-native SaaS products
Remember November 2022? ChatGPT launched and shocked the world. AI was already a thing everyone was discussing in tech companies, but it was not really user-facing. With the launch of ChatGPT, AI became tangible and something everyone tested.
A few months/years later, a vast majority of non-AI-native SaaS products have integrated AI features into their products, sometimes only for the buzz, and sometimes with real utility.
But it’s now common for SaaS products to have at least one key AI feature improving productivity and/or software capabilities.
We’re not anymore in the “Should we build an AI feature?” era, but more in a “What do we build, why, and how do we monetize it?” era.
As the topic is pretty new, it’s hard to find resources on how to monetize those features. Is it a new revenue source? Or only something that we need to integrate into our pricing packages?
There are different models and parameters to take into account to build the best monetization strategy. Let’s dive in!
The 4 AI Features Monetization Models
A/ AI as an Add-on
AI features are sold separately, on top of the standard subscription plans, usually per seat and month/year.
👀 Example: Notion sells “Notion AI” as an add-on → €7.50/member/month (billed annually)
✅ Pros
You isolate the AI feature(s) perceived value and willingness to pay from the rest of your pricing. It’s easier to run tests, as it does not impact your classic pricing plans.
Easy to change the price of the AI bundle.
Perfect if only a small part of your users need your AI feature.
❌ Cons
Buying an add-on creates additional friction, potentially impacting conversion.
It adds complexity to your pricing compared to a single package.
It might feel weird to monetize it as an add-on if your core use case relies on AI.
This model is for me the perfect way to test AI monetization in many cases. And can help to transition to another model by limiting risks.
B/ AI included in high-tiers/paid plans
AI features are bundled into Pro, Business, or Enterprise tiers, with or without a price increase compared to the initial price. Free users can’t enjoy AI features.
👀 Example: Lemlist (cold outreach tool) AI features are gated for non-paying users
✅ Pros
Clear differentiation of free features VS paid features.
Creates a FOMO: people want to try AI features, potentially increasing willingness to pay and then free-to-paid conversion.
Limits AI infrastructure costs to paying users only.
❌ Cons
Free users can’t enjoy AI features, so it reduces adoption and can decrease the perceived value of the tool for them.
Depending on how AI is a core feature of your product, it might be an error to limit it to only paid users.
This model is great if AI features are not the core value of your product, but a nice-to-have that can motivate users to upgrade to a paid plan.
C/ AI in all pricing plans (even free)
AI is made available to everyone, including the freemium users. Depending on the plan, users might not be able to access all AI features, or only with limited usage.
👀 Example: Claap (meeting recording tool) AI assistant features are included in all plans, even the Free one (Basic), with different features gated and AI models
✅ Pros
Accelerates adoption, as AI features are provided by default.
Great to differentiate from competitors who gate similar features.
Easier to communicate on AI features, as all users can try them, no matter their plan.
❌ Cons
It’s hard to measure if AI features are a real monetization driver, as they are cannibalized by other features.
AI infrastructure costs are high, so it can be risky to provide them to all users if it’s not a real revenue driver.
If infrastructure costs are controlled and you managed to prove with a test that AI features are a real driver of acquisition or monetization, then it could be a great model. Otherwise, I would not recommend it.
D/ Pay-as-you-go/credit-based
Users get limited AI usage (eg. credits, prompts), then are completely limited or pay for extra usage. Often, for non AI-native companies, this model is combined with one of the 3 strategies presented above.
👀 Example: Canva AI features are included in classic plans, but limited based on a number of uses per month
✅ Pros
Pricing is aligned with value: users pay for their real usage.
Even with limited credits, people can start adopting AI features, then get frustrated, and buy more credits or upgrade to a more expensive plan to benefit from more usage.
❌ Cons
As it’s generally not a standalone model, it can create complex pricing bundles. Different features, different levels of usage… That can be hard to clearly compare which plan you should go for as a user.
AI costs need to be anticipated properly to create usage limits that both make sense for your business and your users.
Being limited can frustrate users and impact churn.
If your AI features can be monetized as consumption units, this model can make sense. Doing tests and user research is key for validating if a usage-based model is fitting your user needs, and what the “right” usage limits are.
A framework to choose the right AI monetization strategy
The worst thing to do when monetizing AI features is to choose a model by default, without conducting a discovery process, and, eventually, testing different models.
At the highest level, you can use the following framework to start evaluating which model might be best for your product:
But it’s key to go deeper and answer the 4 following questions:
Are AI Features core or complementary?
Is AI essential to the product value proposition?
Or is it a power-up for specific users/teams?
Core → Consider bundling in higher tiers or base plans.
Complementary → Add-on or usage-based might work better.
What’s the adoption potential?
Will >80% of your active users use AI features on a daily or weekly basis?
Or is it only used sporadically, or by some power users?
High adoption → Making AI features available to everyone might be best.
Low adoption → Add-on or usage-based is again probably better.
What’s the willingness to pay for AI features?
Is AI just a nice-to-have, or a real growth lever to increase conversion and revenue?
Will AI features help you close more deals? Increase free-to-paid conversion? Drive expansion?
High willingness to pay → Sell it as an add-on or include it in high tiers to increase conversion.
Low willingness to pay → Include it in some paid plans, but not as an add-on or as credits.
What’s the cost of integrating AI features into your product?
AI infrastructure costs can be huge, especially if reach and usage are high.
How many users will use AI features?
Do you rely on an external model (eg, OpenAI) you need to pay per query to, or is your AI built in-house?
Costs are low → You can offer AI features to all users, or at least paying users, but without usage limits.
Costs are high → Limiting usage with credits and/or only including AI in paid plans is best.
Wrap-up
If you’re a non AI-native SaaS company, integrating AI features is, in many cases, no longer a nice-to-have. New AI players are emerging every day, and some of them are on the path to disrupt industries. Users love AI features (when they’re useful; I’m of course not saying to build AI features without any rationale behind them).
The challenge is to make the most of your AI features, and it happens partially through monetization.
As for every monetization project, the best practices apply:
Explore Beta programs/limited rollouts before going all-in
Conduct a product discovery to understand willingness to pay, usage, etc.
Define a clear success metric when monetizing those features.
Consider testing different models to reduce risks and maximize AI features’ monetization potential.
That’s all for today. See you next week folks 👋
Very impressive content PJ.
Nice post Pierre-John! One more to add - BYOK (Bring Your Own Key).