What the Best AI Companies Have Learned About Pricing

In some of our recent conversations, we've found it eye opening how honest pricing leaders are about the uncertainty in this space.

There's a lot of anxiety in pricing right now. The stakes are high, the risk is real, and nobody knows exactly what north star they're driving toward. Nobody knows if their pricing strategy is the right one, or the best one, if they're taking on too much risk or leaving too much money on the table.

So recently, we embarked on a listening tour. We gathered some of the most forward-thinking leaders in AI pricing and asked them: What does good pricing look like?

We just presented those findings in a talk at this year's AI council, and it was incredible to see the number of leaders who resonated with these points. So we're sharing the recap here for any pricing leader to explore.

Pricing now ships like product, but what's the roadmap?

If you've spent time in the pricing space, you've heard this adage. But more and more, we find that companies often feel like they're flying blind on putting this into practice. It's not enough to just treat pricing as an experiment, it needs to be thoughtful, purposeful and de-risked.

The arc that emerged was that teams are thinking about pricing on 4 levels, each with a different timeline for consideration:

• On launch: Companies tend to prioritize adoption--freemium models, trial credit blocks, self serve availability with simple value metrics. The focus is all on getting users to adopt.
• Weeks: Most companies we spoke to are meeting weekly or biweekly to discuss their costs and make point adjustments to their self serve and entry level plans. The goal is to continually find ways to incrementally charge for new capabilities or better cover foundation model costs.
• Months: Most companies are meeting regularly to strategize how they are converting self serve or small customers into longer term growth. How do you lock in renewals and upsells? How are you ensuring that you're getting paid for your differentiators?
• Years: Many maturing companies are also roadmapping their longterm monetization goals. They're mapping out where their product is headed, the revenue model that they believe will win, and planning incremental steps to get there.

Growth and monetization are often conflated--but they're really different things. A lot of companies treat growth as a customer acquisition problem, or a near-term ARR problem. But real monetization is about understanding how each feature your build is driving incremental revenue, and ensuring that you're building for the users who will pay.

The levers companies focus on to optimize pricing

In addition to thinking on multiple timelines, each company we spoke to is also thinking about pricing on several levels. They're optimizing between cost levers and commercial levers to simultaneously cover the cost of scaling while encouraging their customers to grow alongside them.

Knowing when and how to apply each of these levers was a big topic of conversation. Companies feel an anxiety between delivering a best in class product to customers, and being able to communicate why their premium costs are translating to ROI.

More and more, we're hearing the most mature companies actually separate the conversation about cost from the one about value. Because in usage-based, you grow when your customers grow. But making sure they never feel penalized for that growth, and avoiding making the conversation simply feel like a margins and markups negotiation is essential for long term partnership.

What we found is that finding ways to communicate that value in the commercials, to signal partnership as customers scale, is foundational to any monetization strategy.

Finding your pricing north star means deciding what signals matter

One thing that came up repeatedly in the tour was how differently teams were approaching the question of signal. Some were tracking churn religiously and calling that their pricing read. Others were doing interviews but had no quantitative baseline to anchor them. Very few had a structured way to think about signal across the full lifecycle of a pricing change.

The framework we compiled breaks signals into two dimensions: whether you're gathering them before or after a pricing change, and whether they're qualitative or quantitative. Each quadrant gives you different information, and the teams that navigate pricing changes well are usually drawing from all four rather than over-relying on one.

We wrote a deeper breakdown on the types of pricing signal that matter and how to choose the ones that apply to your company here.

Pricing lessons from the front lines

Three conversations stood out to us over the course of this listening tour. IN each one, a pricing leader described a difficult situation they had experienced with evaluating price changes, and the steps they took to resolve it.

Each one came with a hard-won lesson about how to get the nuanced details of pricing right, beyond simply picking a model and iterating.

Case Study: Why friction isn't always a bad thing

Background: An AI + SaaS company with strong product-led traction wanted to monetize some of their new features.

The Signals: They began with social listening to determine how price sensitive their customers were. Was the user base generally happy? Were they perceived as cheap or expensive in their market?

The Outcomes:

• First, there was a lot of discussion. Some users were upset at the price increase.
• Other users began to explore the new capabilities, and feedback was strong. Discussion began to shift toward these new features vs the price change.
• Overall, user growth continued--because the offering was still valuable and differentiated.

The lessons:

1. The bark is often worse than the bite. If everyone is happy, it's actually a possible signal you're pricing too low
2. Customers care about value, not raw price point. If your product is meaningfully differentiated, you might have more pricing power than you think.
3. How you message a price change matters: Are you pairing it with the launch of new capabilities? Giving users a reason to stay excited? Or simply asking them to accept a new value equation?

One of the scariest parts of price evolution is standing on the value of your differentiated features and charging for where you really drive value. But if the value is genuinely there, customers often broadly accept price increases, even if the initial signal is friction.

Case study: Price for loyalty earlier, not later

Background: An AI-native company realized they could no longer sustain the costs of their entry level price tier. They were simply giving away too much for free and subsidizing too much token consumption under the pricing model. So they set out to increase prices, expecting some churn.

The signals: Revenue pressure was their primary driver, with churn impacts as their lead indicator of success.

The outcomes:

• They began by sharing comms about the change well in advance, and offered existing users blocks of free credits to ease the transition
• Customers still vocalized their anger, and the hit to churn was bigger than anticipated, with many users switching to comparable but cheaper tools
• Churn did not recover as quickly as they had hoped, leading to some tough conversations

The learnings:

1. The churned users cared more about free tokens than the value of their product, and when the subsidies went away, so did they.
2. That wasn't actually necessarily a bad thing: The company was able to use this as a moment to focus on their real ICP, the customers who saw their product as differentiated and valuable, and not just a free token machine.

The Lessons

1. What you charge for is how you communicate value, and if you're not charging for your biggest differentiators, they're being overlooked
2. The complexity of your value metric and your pricing model can help position your product as something more. If your metric is too COGS-aligned, then you may be implicitly training customers to see you as a free token machine.
3. Friction from token tourists isn't the friction signal that will lead you to long term growth, and weeding them out actually helps you narrow down who to listen to and how to build for loyalty.

Understanding and segmenting your customer base is critical for understanding where your real risks are. A power user could be a dedicated champion for your product, but they could also be a token tourist. Sometimes frictionful price changes allow you to find the difference between the two and start building for real loyalty, not just user adoption.

Case Study: How you price determines how you compete, and lower isn't always better

Background: An AI infrastructure company realized that their sales team was struggling to communicate how they were different, and they were being forced to undercut on price, despite being a superior solution on many different axes.

Outcomes:

• Their pricing model optimized for simplicity: just a single value metric. But their competition was going to market on the same metric, meaning that they were being compared apples to apples with tools that didn't build to their same depth.
• Simple metrics may be easy to understand and trust, but they're also very easy to compare, leaving little room for more nuanced conversations.
• Changing to a more multi-dimensional rate card allowed the company to show more value and surface more axes where they were superior.

Lessons:

1. Charging is sometimes what makes a feature or experience premium. If it's not on the rate card, it's not always getting noticed in the value conversation
2. If you can't charge for your differentiators, it might be a signal that you're not building for the right kind of value. Consider rethinking your lens for what functionality is tablestakes and what's meaningfully new to the market.

Being brave enough to charge for your differentiators is initially uncomfortable for many companies. But it's a strong signal that you're building in the right direction. It also adds value levers to a renewal conversation--if your only value metric is a translation of your COGS, then you'll always be negotiating on margins. But adding support, a la carte features, and other dimensions gives your teams the flexibility to cover costs while making it sustainable for customers to scale up usage.

Lessons for pricing leaders

To discuss your pricing challenges, get in touch with our team.