How the Leading AI B2B SaaS Applications are Monetizing

Part 1 of our Show Me the Money: How to Monetize AI [3 Part Series]

In the past two years, AI has already dramatically changed the tech landscape. New AI products and applications have flooded the market, with funding to AI companies doubling quarter-over-quarter to $234B in Q2 2024 [1].

While many AI products haven't fundamentally changed the value that software delivers, some have. These select few are redefining how software creates value for users. How AI apps deliver value is evolving, and will alter how they are ultimately able to monetize down the line once user growth starts plateauing and sustainable ARR becomes a focus. AI undeniably boosts productivity, but how can teams quantify this in a way that passes the scrutiny of CFOs? Are AI apps creating entirely new work products or replacing whole roles? Regardless, it's time we start thinking about how AI is going to make money to justify the lofty valuations that have hit the market.

SaaS Pricing Models 101

Picture a product team that is excited about a new AI product or feature they have built. They might have already believed they have product-market fit after releasing well-received alpha and beta versions. Now it's time to think about product-market-GTM fit and selling this shiny new thing.

SaaS products---whether consumer or B2B---are typically monetized in one of three ways:

Standalone: AI capabilities anchor a new, standalone line of business, often helping established players reach new customers.
Bundled in existing plans: AI capabilities are weaved into existing pricing tiers to drive differentiation and/or boost user engagement and retention. They can be offered in all current plans to enhance product value (with or without a price increase), or select premium plans to drive expansion ARR opportunities for lower-tier customers to upgrade.
Add-on: AI capabilities are sold as a complement to existing plans. This approach expands existing customer ARR for customers that are interested, but avoids over-bloating the product with niche features that not all customer want.

For a new company going to market with its first product, the default answer is standalone. However, if you already have products in-market, all three options are viable. So, how do you determine the best approach to monetizing your AI capabilities?

While there's merit to considering cost-based and competitor-based pricing approaches, the best way to monetize AI will ultimately still be a value-based approach --- in the long-run, you must provide your customer with product value equivalent to what they pay. A good starting point is understanding your customer's willingness-to-pay vs. their interest (or initial beta adoption) in your AI capabilities. Below is the framework I often use when I work with founders on how they should think about growing and monetizing their new products. You don't want to give away all of your value-add features, but you also don't want to try to nickel-and-dime your customers who may love and advocate for your product.

AI capabilities with high interest and high willingness-to-pay are able to anchor new, standalone lines of businesses with monetization separate from existing product lines. Those with high interest but low willingness-to-pay, however, may be better suited bundled in current plans to drive better engagement and retention with current products via indirect monetization. Finally, products and features with low interest but high willingness-to-pay are prime targets for add-on monetization, particularly if the niche addressable market is lower compared to your main product line.

This framework provides a great starting point for hypotheses, but should also be considered in conjunction with your overall go-to-market strategy, financial considerations, and customer segmentation and target ICP. We'll dive more into SaaS pricing models in Part 2 of our series, when we consider alternatives to traditional SaaS monetization models for AI. If you're a founder, you won't want to miss this!

Monetization Trends for the 100 AI B2B SaaS Applications

AI products are evolving rapidly, which means pricing and product decisions are also rapidly iterating. We took a look at the leading 100 B2B SaaS with AI capabilities and valuations >$1B in the application layer to understand how they are monetizing today [2]. These companies use various business models, including freemium, subscription, usage-based pricing, and custom enterprise pricing. Here's what we learned:

Most companies bundle AI features in their current plans, but many are experimenting with add-ons. Over 50% of the companies, such as Asana and Canva, we analyzed bundled AI features into their existing plans, particularly in higher tiers to encourage plan upgrades from existing customers. This approach allows them to learn about feature adoption on untested AI capabilities and quickly iterate on getting to the best pricing model. About 25% of the companies we analyzed are using hybrid models, offering basic AI features for free while charging for advanced capabilities as add-ons to boost ARPU.
Traditional SaaS verticals use AI as a key differentiator to drive retention and open new revenue streams, while modern SaaS verticals use AI for acquisition. In established markets focused on automating legacy business functions, companies like CrowdStrike's Falcon AI and Workiva's AI-powered governance tools are using AI not just as an enhancement but also as a primary driver of product innovation. Modern SaaS players, on the other hand, focus on showcasing AI in free trials to acquire new users --- over 90% are offering a glimpse of AI features in their free trials and freemium.
Larger companies are less willing to give away AI capabilities for free, but are more willing to throw multiple darts on the board to get to the right pricing models. There's a clear trend that the larger the market cap of these B2B SaaS companies, the smaller the percentage of them are putting AI capabilities into free trials or freemium versions. In fact, apps with >$50B market cap overwhelmingly do not offer free versions of their AI capabilities, with <25% with free trials. Their advanced AI features are either paywalled completely in their higher tiers with no free trials, or are sold as another plan or add-on. These same larger companies also have a wider distribution of different pricing models for AI across standalone, hybrid, bundled in current plans, and add-ons.
The age-old per user value metric is still dominant, but that's quickly changing. We're seeing more companies leverage both disruptive value metrics and hybrid approaches. AI SaaS applications in cloud infrastructure and data management are leaning into usage-based AI pricing models; this allows them to scale pricing with the volume of data processed or stored, aligning with customer needs as they grow. By tying pricing to metrics like data volume or API calls, these companies can capture a proportional share of the value their AI capabilities create, incentivizing growth.
Pricing models are evolving with sophistication and becoming more nuanced, as companies start adopting outcome-based or work unit-based pricing for AI. As AI faces existential ROI questions, more companies are experimenting with successful work unit as a new value metric rather than a per user basis. Salesforce is charging $2 per conversation for Agentforce. Zendesk is pricing per successful autonomous resolution with Zendesk AI. Intercom customers pay $0.99 per successful AI resolution. The list goes on. We expect these pricing models to become even more common as the old per user model is now being questioned as customers are likely buying far more seats than they need as the true ROI of AI leans towards delivering actual work units rather than augmenting productivity.
Revenue growth is uncorrelated with type of pricing model, for now. How you monetize will undoubtedly have a direct impact on revenue growth, but people overestimate how long it takes to create a sustainable growth model. There's too much evolving with AI capabilities to say that one particularly pricing model drives more growth, and the data shows that. Ultimately, each company may have different outcomes with different pricing models based on its core features and ICP. But, the data does hint that those companies with lower revenue growth thus far may be placing big bets on AI to anchor a new standalone business to revive growth. Whether that works or not remains to be seen.

Final Thoughts

It still feels like we're in the early days of AI applications. New startups have yet to reach true product-market fit yet and legacy SaaS companies are trying not to fall behind while testing monetization hypotheses to see what sticks. We know that there's market demand for AI apps and its quickly commoditizing SaaS, but it's also just hard to sell AI right now. It's expensive and the true ROI is still a mystery. Most AI apps are unable to answer the top questions we've heard buyers ask today:

What improvement can we expect from this AI app compared to the SaaS solution I already have today?
How much better is this AI app vs. its peers?
Am I replacing existing software spend and buying a better version or replacing actual human labor?

The evolving SaaS value metric from per user monetization to something else is going to be the key to defining how companies are really going to make money in AI and how investors should evaluate AI app business models.

For now, here are the top takeaways we have for the prepared founders thinking a few steps ahead on monetizing their AI capabilities down the line:

Hybrid monetization models can be powerful to tap predictable revenue from subscriptions while scaling, and also gather data on ICPs that would directly pay for AI in the early days
Early and iterative pricing experimentation is key. Be prepared to shift pricing strategies based on customer feedback and be creative in testing what works. This means a flexible billing platform and payments infrastructure to enable this
Move beyond the traditional per user pricing models, particularly if your AI product is highly scalable and consider tying it to usage (e.g. data processed, API calls) or tangible business outcomes that buyers would relate to (e.g. successful actions, resolutions)
Build toward long-term revenue streams. Consider how your pricing model can evolve once the initial excitement of AI wears off and customers demand measurable ROI
Talk to your customers! Gather insights from early adoptions to refine your monetization strategy. Don't assume that just because users adopt your product, they're willing to pay for it in its current form. Understand what drives their willingness-to-pay and tailor your pricing accordingly.

In Part 2 of our series, we're going to dive next into value metrics and alternative monetization models that AI apps can consider.

Notes

Crunchbase, https://news.crunchbase.com/data/global-funding-ai-biotech-h-2024/
Internal database, based on leading 100 B2B SaaS companies with AI capabilities in the application layer with market cap >$1 billion.