With the rapid development of AI technology, the pricing model of software companies is undergoing a major transformation. By analyzing the pricing data of 240 AI software companies, this paper reveals five key trends: seat pricing and fixed price models are facing challenges, hybrid pricing models have become mainstream, multiple pricing strategies have their own advantages and disadvantages, the resulting pricing model needs to focus on four major issues, and price transparency is not suitable for all enterprises.
While AI is profoundly changing the software industry, it also brings a problem:
While AI capabilities are powerful, traditional pricing methods are failing due to value misalignment and cost pressures.
Against this backdrop, software companies are more in demand than ever for new disruptive pricing models. From now on, this trend is progressing faster than everyone thinks.
Recently, foreign technology author Kyle Poyar collected data on more than 240 software companies with annual recurring revenue (ARR) between $1 million and $20 million, selling hybrid SaaS and AI products.
Using data from these 240 software companies, Kyle Poyar came up with 5 trends about AI pricing:
- Traditional seat pricing and fixed-price models face challenges, and hybrid pricing models have become mainstream.
- When hybrid pricing became a trend, some new pricing combinations emerged.
- Outcome-based pricing is good, but it doesn’t work in the short term in most markets.
- The value of price transparency may be overestimated.
- Pricing models are still changing rapidly, but most companies aren’t ready for it.
01 Seat pricing and fixed price models are facing challenges
Twelve months ago, software pricing was mainly based on seat charging and fixed-rate subscription models.
These patterns provide predictability in prices and promise long-lasting recurring revenue (ARR). However, these models are increasingly under threat due to value misalignment and cost pressures, especially for AI-native products.
How can product managers do a good job in B-end digitalization?
All walks of life have taken advantage of the ride-hail of digital transformation and achieved the rapid development of the industry. Since B-end products are products that provide services for enterprises, how should enterprises ride the digital ride?
View details >
Flat fee and seat-based pricing are being replaced by hybrid pricing, which is a combination of subscription and usage. This trend can also be seen in the data:
- Over the past 12 months, the share of fixed-fee subscription models has dropped from 29% to 22%.
- The share of seat-based pricing has decreased from 21% to 15%.
- The proportion of hybrid pricing models has increased from 27% to 41%.
Another trend is that AI and software are becoming more and more closely integrated.
More than half of respondents (53%) said they incorporate AI capabilities into their core software products. Only 20% of respondents said they do not offer any AI features. Even fewer respondents (16%) said AI is primarily sold as a standalone product or add-on.
One of the changes that AI brings to software is that software provides more and more value. Customers may need less labor and more AI.
- Alphabet says that more than 30% of its code is currently generated by AI.
- Microsoft’s CTO expects that by 2030, 95% of code will be generated by AI.
- With 60 employees, Cursor grew to $200 million in average annual recurring revenue (ARR) and more than $3 million in average revenue per employee.
- Klarna said its ARR per employee soared from $575,000 to $1 million due to improved AI efficiency.
But on the other hand, the cost of delivering AI features is real and is becoming a key factor in pricing. Survey participants believe that internal costs and profits are the most important factors in pricing AI features.
02 Hybrid pricing model has become the mainstream pricing model
In the SaaS era, most software companies’ pricing strategies mainly refer to Salesforce or Slack. In the era of AI, their pricing strategy was inspired by Clay.
Clay uses a hybrid pricing model that meets user needs through a variety of pricing methods while keeping pricing relatively simple, such as more features (subscription packages) and more usage (credits).
Instead of offering a significant discount on the annual plan, Clay offers a small discount (10%) and allows customers to get all their credits at once. Unused points can be carried over to the next month (up to 2 times), which is convenient for customers and increases user stickiness.
Recently, some startups and large enterprises have introduced a similar hybrid model. For example, I’ve been following monday.com (all paid plans now offer 500 AI credits per month), Salesforce’s Agentforce (which added an elastic credit model in May), Atlassian, and many others.
Hybrid pricing is a natural evolution of seat or flat-fee subscription models. I think there are four reasons why hybrid pricing is so popular:
- It did not have much impact on the original pricing system. Hybrid pricing doesn’t have to look the other way – it can fit into existing seat-based and subscription-based models.
- A more natural path to sales. It creates a natural upsell path for customers to try out new products “for free” and then monetize as usage grows.
- Considerable profit margins. By limiting usage, companies can control costs and minimize the risk to unprofitable customers.
- Relatively predictable. By following traditional pricing models, buyers can estimate costs and control their spending.
03 Seven common pricing strategies
As more and more AI products move to hybrid pricing, a new challenge arises: there seem to be countless ways to build hybrid pricing, but not every one is suitable.
Here, the author shares some common pricing methods, along with their pros and cons.
First, Pay As You Go (PAYG).This is not really a blending mode, and may have good results early on. Pay-as-you-go means no commitment, complete flexibility. This model works best when customers can reimburse expenses or include them in the operating budget. Otherwise, enterprise procurement should be careful!
Second, there is a capped pay-as-you-go (PAYG).This model provides peace of mind to buyers by limiting potential usage/spending. This pattern is increasingly common in outcome-based patterns, as the results are previously unknown.
Third, usage-based packages.Customer promises a certain amount of usage or package; This is usually “while supplies last”. Packages include a variety of sub-modes, including high billing, where customers immediately enter overbilling mode if usage exceeds the plan cap, or decreasing mode, where usage can be used flexibly, just like gift cards. Concerns about overuse and fluctuating usage can lead to overselling for salespeople and overbuying for customers.
Fourth, platform fees plus usage.Charging platform fees helps lock in customers while giving them access to premium features, premium support, and more. This approach is effective when pricing metrics are commoditized (e.g., SMS, calculation, storage) or do not reflect the full value of the product. Vendors can advertise affordable prices but need to be compensated by platform fees.
Fifth, platform fees (including usage) plus additional usage fees.This model, also known as the three-part tariff model, has a higher subscription fee but includes a certain amount of “free” usage fees. Offering a minimum usage helps attract customers and often stimulates them to increase their overall spending.
Sixth, adaptive fixed rate.In this model, the customer commits to choosing a usage-based tier but can use the product as much as they want for the duration of the contract without incurring overage fees or requiring upgrades. When a contract is renewed, its tier is adjusted up or down based on actual usage. Adaptive flat rates are predictable for customers while also encouraging them to increase their usage over time (the downside is that if usage drops, you’ll still have to pay for it!). ）。
Seventh, platform fees plus success bonuses.In this model, pricing is presented in the form of a more traditional subscription fee. If the return on return (ROI) received by the client is higher than expected, they will be charged an additional bonus or commission.
04 Result pricing, 4 issues that must be paid attention to
5% of respondents said their current primary pricing model is results-based. However, 25% of respondents said they don’t expect to shift their pricing model to an outcome-based model until 2028.
Early adopters of outcome-based pricing – such as the previously mentioned Intercom – are paving the way for the market.
Some of these so-called “outcome-based” models should be more accurately called “work-based” pricing (e.g., EvenUp, Casemark). Others are true “success-based” pricing models, where suppliers take a share of the customer’s revenue (e.g., Chargeflow, Flycode, and AirHelp).
AirHelp, for example, charges a 35% success fee when it wins a customer compensation for flight delays or cancellations.
When an AI agent is positioned as a role that “performs a task,” it makes sense to price it based on the amount of work it does (or the benefits associated with that work). From a marketing point of view, this model has been very successful.
It sends a strong signal that you are confident in your product and willing to give it your all. At the same time, this also prompts suppliers to continuously invest resources to improve product effectiveness, thereby bringing more practical results to customers.
However, there is also a problem behind this model that cannot be ignored, which I call the CAMP framework. To achieve outcome-based pricing, businesses must have the following four elements (CAMP):
1) Consistency: Do all customers value the same results? Or is it that different customers need different outcomes, leading to the need for customized results, which in turn leads to a large number of customized contracts.
2) Attribution: Can you convince customers to attribute the results achieved to your product? Or do they think that they are mainly on their own and that your product is only helping a little?
Outcome attribution is one of the most difficult issues in outcome-based pricing. If customers can’t clearly see the role your product is playing in driving results, they’re less likely to pay for it, especially if it’s a result share.
3) Measurability: Can you measure and report on these outcomes in real time? Or do you need to rely on customer reports, A/B testing, and/or proofs of concept to confirm results?
If results cannot be measured accurately and in a timely manner, it will be difficult to establish transparent billing mechanisms and trust relationships. Ideally, the system should have the ability to automatically track key metrics and show customers clear evidence of value.
4) Predictability: Can you predict the results of your product with some accuracy? Or are results very different between different customers?
If the results are volatile and unpredictable, the business will face significant financial risks. For example, some customers may receive extremely high returns, while others have little to no improvement. This makes it difficult to standardize the pricing model and also makes it difficult to sell.
05 Price transparency may be overestimated
The practice of hiding pricing information was once considered a legacy of the 90s and 2000s. After all, savvy buyers today will do their research online (or ask their peers) and will likely find the price information they want.
Tools like Vendr even show what others actually pay through a free Chrome plugin.
Open pricing allows you to capture this segment of buyer demand (and associated search traffic) while maintaining narrative power (i.e., proactively defining the customer’s understanding of the value of your product). Plus, it filters out unqualified buyers so they don’t waste your team’s time.
However, the reality is that despite the advantages of transparent pricing, many companies do not fully adopt it.
This may involve complex pricing structures, differentiated pricing strategies, or fears that price will become the focus of competition and weaken the value proposition. Therefore, although the trend of transparency seems “inevitable”, there are still many challenges and concerns in actual implementation.
The reality is: things don’t go exactly as expected.
Businesses with an average annual contract value (ACV) of less than $5,000, as well as companies with a product-driven growth (PLG) model, often do make pricing information public on their websites. But for other businesses, this is not so common.
My opinion: Many software companies, especially start-up companies and AI-related companies, have not yet fully clarified their pricing strategies. Once the price is made public, it becomes much more difficult to adjust it later – because it can easily confuse consumers and even lose trust.
In addition, as pricing models become more complex (e.g., hybrid pricing with AI credits), buyers don’t necessarily believe that the price they see on the website is what they will eventually pay.
Naturally, they ask a series of questions: Is there a cap on use? How to calculate excess fees? What features are available for an additional fee? Wait a minute.
When complexity rises, buyers prefer to communicate with real people rather than relying solely on a price list on a web page.
06 AI pricing is changing rapidly, and most people are not prepared enough
The rapid development of AI technology has made it no longer feasible to blindly follow the existing pricing model. (In fact, three-quarters of software companies adjusted their pricing strategies last year.) ）
As pricing decisions become increasingly strategic and complex, businesses need to dedicate resources to pricing efforts. There is a lot of practical work behind this, including a deep understanding of cost structures, competitor dynamics, and customer perceived value.
However, most companies still fall short in two areas:
- Personnel Gaps: Lack of talent with professional pricing analysis, value modeling, and market insights;
- Legacy Tooling: Still relying on traditional Excel sheets or outdated systems that cannot support real-time, data-driven pricing decisions.
In other words, while pricing is becoming more important, many companies do not have the ability to match it to support this strategic transformation.
In the early stages of a company’s development, pricing is almost always determined directly by the founder or CEO. Subsequently, as the company expanded, pricing gradually became a “hot potato”, shirking back and forth between multiple departments such as sales, product, marketing, finance and operations.
Be especially wary of falling into so-called pricing “no-man’s land” – which usually happens when annual recurring revenue (ARR) is between $5 million and $20 million. At this stage, the “pat your head” decision-making method in the start-up period is no longer applicable, but the formal pricing mechanism and responsible person have not yet been established, resulting in a lack of clear ownership and strategic direction in the pricing strategy.
07 Summary
I’m still optimistic about usage-based and hybrid pricing models, but I’m increasingly inclined to think that they’re just transitioning to work-based and outcome-based pricing.
This is actually part of a broader evolution of the entire software industry from “owning” to “renting” and then to “on-demand”:
- From on-premise to subscription-based (SaaS);
- and then move from the subscription system to the payment model based on actual use;
- Each evolution is reducing upfront costs for customers and making software more accessible;
- At the same time, it also shifts the risk from the buyer to the supplier, forcing the supplier to be truly responsible for the customer’s results;
If we can achieve this vision—a truly mature outcome-based pricing model—it will revolutionize the way software companies operate. Every department in the company will operate around one core goal: to help customers achieve their goals.
