👋 Hey, I’m Lenny and welcome to a 🔒 subscriber-only edition 🔒 of my weekly newsletter. Each week I tackle reader questions about building product, driving growth, and accelerating your career. For more: Best of Lenny’s Newsletter | Hire your next product leader | Podcast | Lennybot | Swag

Q: What trends are you seeing in how incumbents are pricing AI features and products? Have you seen any innovation in how companies charge?

Pricing and AI: individually, two of the most mysterious topics for product builders. Put them together, and things get dizzying.

To answer your questions in the most comprehensive way possible, I teamed up with Palle Broe to analyze how leading tech companies are approaching AI pricing and, from that, create a framework to help you make decisions about how to price your own AI products and features.

Palle Broe was part of the early team building Uber in San Francisco and the U.K. and then spent five years across various operational, strategy, and product roles at the company. He led pricing strategy while at Uber (B2C) and then at Templafy (B2B SaaS), and has advised more than 20 fast-growing tech startups (seed to Series D) on monetization strategy, creating better packaging, cutting back on discounts, and building stronger ROI and business cases. I’m excited to offer you his insights on pricing for AI features and products, which looks very different than it does for other technologies.

For more from Palle, make sure to subscribe to his newsletter, Rule of Thumb, where he provides tangible pricing advice to operators, and follow him on LinkedIn.

AI features and products present brand-new pricing challenges for companies. I’ve spent the past decade working on monetization strategies for places like Uber and Templafy, as well as advising more than 20 tech companies on their approaches, and what I’m seeing around AI products is very different from older technologies. Unlike with traditional SaaS products, companies looking to integrate AI products and features need to consider the real, underlying costs of generative AI compute and the intense competitive pressure in the AI market now, while also focusing on adoption and new business models. To ensure long-term ROI, companies have always needed to think carefully about how and when to monetize, but AI requires even deeper analysis.

During Alphabet’s 2024 Q2 earnings call, many questions from investors addressed the ROI of the company’s $12 billion AI investments—a real shift from the previous quarter’s call. It’s clear that investor focus is changing from pure adoption to how big tech is going to be monetizing their innovations. 

This left me wondering: How do tech companies monetize their new AI features today? And what can we learn from that data?

To answer, I investigated the pricing and bundling strategies of 44 leading tech incumbents. I focused on the “application layer”—companies that are building end-user products (e.g. Figma)—rather than base models (OpenAI’s LLM) or infrastructure (e.g. Azure). We reviewed public data for pricing models, value metrics, bundling, and free versions to identify current trends. Based on that data and my own experience in pricing, I’ve put together a framework for you to help make strategic decisions for your own AI products and features.

1. Direct and indirect monetization strategies

Broadly speaking, there are two methods that companies can use to monetize AI features: direct and indirect monetization. Direct monetization involves charging for the AI feature directly, or increasing the price of your product after adding the new AI feature. Conversely, indirect monetization integrates the AI feature into an existing bundle without altering the price, or offers the feature on its own at no additional cost.

Below is an overview of the five high-level monetization strategies we are seeing right now for tech companies launching AI features and products.

What companies are doing now

The predominant strategy we saw in the data was to bundle the AI features into existing packages (59% of companies chose this path). This approach allows users with current subscriptions to benefit from AI capabilities. In some instances, this integration is accompanied by a price increase or usage-based pricing for the AI feature—making it a direct strategy. In others, it is added without altering the existing pricing structure—an indirect approach. Adding AI features to an existing bundle can be used as an interim strategy to launch quickly and before thinking more carefully about monetization as more data is gathered around usage of the AI feature. 

The second most common approach is a direct strategy: offering AI features as an add-on with a distinct price tag (23% of companies chose this path). The add-on strategy is the “purest” form of direct monetization and will provide you with the cleanest data in terms of adoption and monetization. Also, the ability to track the direct impact of your AI feature will enable you to understand willingness to pay and can provide important feedback to the roadmap and product development. 

Additionally, some companies (18% of those reviewed), particularly those with large language models (LLMs), have developed standalone AI products available for separate purchase, independent of any existing subscriptions.

How you can choose between direct and indirect monetization

There’s a reason the majority of the companies we looked at used a direct monetization strategy. As a rule of thumb, I believe direct is best, either offering the new AI feature/product as an add-on or bundling it in the existing plan with a price increase (or usage-based component). And this seems to be the path many of the largest tech companies took. It’s critical to understand willingness to pay and the underlying cost structures associated with gen AI, and employing this strategy will enable you to understand both. One of the core issues in applying indirect monetization is that it can be very hard to track and accurately attribute the value from increased retention and upsell.

Direct monetization is likely the right choice if your company has:

1. High variable costs: The variable costs associated with gen AI are significant and cannot be absorbed by indirect revenue gains—for example, the cost associated with compute, bandwidth, data storage and labeling, security and compliance, as well as maintenance and upgrades. LLMs such as ChatGPT, Gemini, and Claude incur very high computing costs, while companies leveraging LLMs such as Airtable pay every time a user uses their AI feature. 

2. Clear customer value: Customers clearly recognize the added value that gen AI features bring them and are thus happy to pay for it—for example, GitHub’s coding Copilot or Intercom’s AI bot, Fin.

An indirect monetization strategy (e.g. including it in a plan without a price increase, or giving it away for free) can be successful when generative AI features significantly boost usage, conversion, or retention of your core product. This results in indirect revenue gains that outweigh the costs for these features—particularly when you have usage-based pricing or when the AI feature greatly increases overall customer conversion or retention. Zoom and Shopify are two examples of companies that have pursued this strategy. Sometimes this is an interim strategy to get user feedback before integrating a price increase once the value to the user is better understood. Making price increases with a large customer base is not an easy task and needs to be handled extremely thoughtfully. In my experience—and in the data—indirect monetization is usually the less attractive choice, particularly long-term.

In reality, very few companies get to choose their monetization strategy in isolation. If a competing company is launching a similar AI feature but choosing an indirect monetization strategy, you will have to take that into account alongside all the other variables above. It might be the right decision to follow suit to ensure competitiveness, but ultimately this will depend on multiple factors not discussed in detail in this article. 

2. Direct monetization deep dive

Next, let’s dig deeper into the three core strategies for direct monetization (add-on, standalone, and included in plan with a price increase) and break down which path might be the best for your company. Typically, the factors to consider are the value your AI feature delivers for the user and your business, the best bundling approach, and the optimal distribution of the AI features across different package tiers.

Step 1: Defining the role AI will play in your product portfolio

Consider the following questions:

1. Will this feature be widely used by a broad audience? Or does it cater to specific personas [y-axis below]? In most companies I have worked with, the benchmark for whether or not to include a feature in a bundle or as an add-on is: If over 70% of users are likely to utilize the feature, it is advisable to bundle it in a standard package. If usage is expected to be below 70%, you need to think carefully if it will make more business sense to include it as an add-on instead—particularly if the feature has high value for a small group of users. 

2. Will a critical mass of people want to pay for this feature [x-axis below]? You want to understand if your feature is a nice-to-have versus a need-to-have. The best way to get this insight is to talk with your users. You can do this via a beta program to get usage data, as well as asking potential customers about their willingness to pay for the feature.

Once you’ve gotten answers to the above, you can place the feature on the map below to identify the best path for bundling your AI feature. As an example, if your AI feature is one that is widely used by users (more than 70%) and adds a lot of value to their work, it’s a “leader” feature in your bundle and could likely result in a price increase. On the other hand, if your AI feature is used by relatively few users in a company (say, 20%) but those users love the feature, it should be an “add-on” to your bundle:

Step 2: Evaluate the three direct strategy options