Saved To My Saved Content
Download Article
Use BCG Answer to summarize this article

This article is a collaboration between BCG, Stripe, and AWS.

Agentic AI is the biggest platform shift since cloud, representing a $3-trillion-plus revenue opportunity for the software industry. This should be good news for SaaS providers, but the market reaction has been sobering. In early 2026, US software stocks lost about $2 trillion in value, on fears that agentic AI would fundamentally undermine their business model. One top-performing tech fund manager told Bloomberg that application software faces an “existential threat.”

Although much of that decline has since reversed and revenue growth in non-AI SaaS companies remains strong, according to Stripe data, the recovery has been uneven and market uncertainty remains high. In our discussions with SaaS CEOs and investors, the same questions continue to be asked: How do we shift our pricing model to ride the AI wave rather than get buried by it? Do we abandon subscriptions and seat-based models altogether? Should we price to token costs?

Answering these questions requires reimagining how to engage customers, structure packaging and pricing, plan go-to-market, measure value, monetize, and organize operations. The result will be a SaaS provider built to keep pace with how AI is reshaping both costs and what customers value.

To create this structure, however, providers should reengage their client base to fully understand what drives value in the AI era. This will inform a new strategy that maps how AI and agentic economics will evolve and how pricing, packaging, and providers’ own cost base can shape that evolution. This strategy can inform a reshaping of the organization to support testing and implementation of a wide range of possible pricing models, including usage-, outcome-, and credit-based models, all of which are rising in prevalence.

Creating this kind of future flexibility is vital. Success will go to companies that recognize we are in a new, fast-moving era and can iteratively experiment, adapt, and implement fresh approaches to monetization. We call this capability “pivot-ready.”

Why AI Is Putting SaaS Pricing Models Under Pressure

SaaS providers that until recently had a well-established and reliable playbook for monetization and business growth now face multiple challenges, some of which predate the early 2026 meltdown:

More fundamentally, customers were already asking tough questions about value. As SaaS platforms have proliferated across enterprises, aggregate spending has become increasingly visible. Inevitably, CFOs have been wanting to see ROI.

That does not mean SaaS players should abandon traditional pricing models and start charging only for usage or outcomes. A complete reset carries significant risk. SaaS revenue growth appears resilient, according to Stripe data. For most SaaS providers, the downside of an abrupt shift to usage- or outcome-based pricing is even greater than the worst-case scenario with existing pricing models—even for providers that face strong competition from disrupters and whose clients are aggressively cutting licenses (“seat compression”). In addition, most SaaS companies lack the organizational capacity to implement usage- or outcome-based pricing.

Nevertheless, SaaS providers (and other software firms) should still urgently become pivot-ready. They should:

Underneath all this is another issue: costs. Traditionally, in at-scale SaaS businesses, these were a background issue. In the AI era, costs vary unpredictably between customers and with the pricing plans of third-party LLM providers. They are also a weak point against AI-native rivals—and, perhaps most important, they are watched closely by investors.

Monthly Newsletter Subscription
Tech + Us: Harness the power of technology and AI

Reengage Customers on What Truly Drives Value

Many SaaS providers cannot meter how their software is used, let alone the value they deliver to customers. Some AI features they have introduced are convenient time savers for individual users but fail to deliver the enterprise-level value that would make customers happy to increase their spending. (See Exhibit 1.)

Chart showing the types of SaaS product features that use agentic AI to drive revenues, according to BCG interviews

This needs to change. To capture the opportunity provided by agentic AI, SaaS providers need to understand and then win their customers all over again. In particular, they should:

Table showing four levels of agentic AI features in SaaS platforms and how they can be monetized, based on BCG analysis

Define How the Pricing and Monetization Game Is Changing

Executives at SaaS providers intuitively know that the rules of pricing have changed. The challenge is in the specifics: understanding which rules have already changed and how quickly they need to respond to changes that are still ahead. They can get clarity by asking three questions.

How much is the market landscape and strategic pricing game changing? Discussions with customers provide key intelligence here. It’s important to understand where customers are spending their “AI wallet”—the budget set aside for AI initiatives—and how workflows are evolving.

For many SaaS players, the core platform will not be eroded, but growth will slow as customer spending shifts to other platforms and incumbents struggle to keep up with growing complexity. Many SaaS companies have historically thrived in what BCG calls “the choice game” (typically good-better-best subscription tiers). The market is evolving, however, and the complexity of AI deployments means potentially shifting to other pricing strategies, such as the cost, custom, and dynamic games. These pricing games require better awareness of how costs scale with AI usage, and greater flexibility and responsiveness.

What should be monetized? Historically, this was clear. Human beings sat in front of a device and accessed the SaaS platform to do their job. But SaaS providers that are pivot-ready can consider three rapid shifts:

How much change does the pricing model need—and when? Seasoned executives and investors have lived through turbulence before; after all, the SaaS business model itself was a highly disruptive force in the early 2000s at a time when software was a “buy once” purchase. Traditional software companies that were not pivot-ready faced a crisis.

Looking at today’s SaaS landscape, it is premature to declare that seat-based pricing is dead. But cutting seats has long been a goal of cost-conscious IT departments, and agentic AI gives them new ways to do this. Three alternatives allow departments to avoid the squeeze: adopting a subscription model more aligned with value (such as an enterprise footprint model), pricing on the basis of tasks or jobs completed, and charging on the basis of operational or financial outcomes. Token-based pricing or infrastructure measures could be used as secondary metrics if cost-to-serve becomes a pressing issue, but we recommend avoiding them as primary metrics. AI token cost curves are like a roller coaster with unpredictable twists and turns; we generally avoid linking revenue to a fairground ride.

The current leading approach, based on experience across BCG, AWS, and Stripe, is the hybrid model, which mixes conventional subscription-based pricing with one of the alternative models listed above. Stripe surveyed more than 2,000 global businesses and found that 49% of high-growth and 57% of hypergrowth companies have adopted hybrid models. And customers are prepared: BCG’s IT Spending Pulse survey and market research across enterprise buyers have confirmed most buyers also prefer the predictability of hybrid and committed-spending models for now.

Building the Infrastructure for Pivot-Ready Pricing

The growth potential of models that incorporate usage is clear, but the shift is harder than many SaaS leaders anticipate. In Stripe survey data, 67% of hypergrowth companies reported that they changed their pricing three or more times in the last two years; for low-growth companies, this was just 22%. The top 100 AI companies on Stripe reach $1 million in annualized revenue in a median of 11.5 months, about 4 months faster than the fastest-growing SaaS companies. Aligning internally on the pricing model and metrics is an essential first step. The companies making the fastest progress treat monetization as core infrastructure, not something to solve in one engineering sprint.

Once aligned, the work begins by defining what the customer is paying for and how to measure that activity, often by building telemetry—collecting the data that will power the new usage-based pricing element. With many new AI capabilities, you can’t monetize what you can’t meter, and the fastest-moving companies capture usage at the event level, decoupled from product code, so pricing inquiries or changes don’t require engineering intervention. The pricing design should also determine whether the new metric is added to an existing subscription or replaces it, and what the company will track to evaluate the model. The investment in billing is vital, and without that alignment, confusion can arise with field teams and customers. An invoice the customer cannot understand erodes trust, risks retention, and creates a nonpayment risk.

Legacy systems, designed for a much simpler world, typically break when faced with complexities in metering (such as turning a given volume of API calls into a dollar figure on an invoice) or portfolio and pricing model management (where products, tiers, discounts, and rates have proliferated). A centralized product catalog with applicable rates reduces engineering overhead and makes pricing experimentation viable.

Without improved pricing architecture, downstream teams that rely on billing information quickly feel the strain. Sales reps are forced to improvise on pricing—and deals slow down—when price sheets and billing tools are disconnected. In finance, the inherent variability of usage-based pricing exposes the limits of traditional financial-planning tools; a shift to processing data continuously will be needed. Without well-constructed monetization infrastructure, teams like these will waste time with manual workarounds.

A high-quality billing architecture can provide additional benefits; for instance, it surfaces insights for the SaaS provider’s marketing and finance teams, as well as insights for customers.

It’s no surprise, then, that leaders often underestimate the upfront resources needed, the roadmap tradeoffs, and the ongoing cost of engineering, infrastructure, and monitoring.

Leaders can return to core questions: Can we ingest and meter high volumes of data, update pricing quickly, calculate invoices accurately across every product and customer, maintain GTM flexibility, and verify that new products are delivering on the R&D thesis? That is not a billing question. It’s an infrastructure question.

To make it work, three shifts are needed:

SaaS Lessons from AI Natives

AI startups are showing remarkable revenue growth. Stripe research shows that AI startups founded in the previous three years have reached $5 million in annualized revenue in a median of just 9 months. In contrast, earlier research showed the top 100 SaaS companies reached this level in a median of 37 months. (See Exhibit 3.)

Data from Stripe showing that recent AI startups are reaching annualized revenue of $5 million in just nine months

These fast-growing AI companies include not only household names such as OpenAI and Anthropic but also specialists in verticals such as legal or health care. Examining these new players yields lessons about how growth models can align with the ways in which a product delivers value. Common patterns:    

Together, these create a systemic challenge to the economics that most SaaS providers experience: to increase revenue, they will also need to increase their head count. AI natives didn’t have to become pivot-ready—they were built that way from day one; incumbents now have to match. 

The urgency is also evident in how quickly young SaaS businesses are now scaling: platforms on Stripe that went live in January 2026 are reaching $1 million in payment volume at a higher rate than any previous cohort over the equivalent period.

Building for What Comes Next: When Machines Become Buyers

There is one other key change. It starts with a foundational question: Why does pricing exist?

The traditional SaaS playbook is rooted in how the human brain processes information and weighs costs against value. Pricing schemes reflect this; for instance, good-better-best is grounded in the human tendency to select the most balanced, “better” tier. Other pricing schemes use price anchoring, creating a premium-priced deluxe package that makes all other tiers appear more affordable to potential customers.

Remove the human, however, and these dynamics lose relevance. AI agents will increasingly have a seat at the table when organizations are choosing SaaS providers. Although they are unlikely to have the final say on the vendor in the near term, they can evaluate and fine-tune thousands of purchase combinations to recommend which vendor or tier combinations generate the most value for the lowest price. In the medium term, they may choose a vendor, manage contracting, and rebalance subscription and usage options with ruthless efficiency every month. These shifts would further erode the cost and revenue predictability that has made the SaaS business model so durable.

An early example of this purchase flow is with PostHog, a software product development and analytics platform. Through Stripe Projects, developers and their agents can provision PostHog services, generate and store credentials, and manage usage and billing from the command-line interface. In just a few months in 2026, PostHog reported more than 200 headless sign-ups, including sign-ups that were agent-driven.

The pace of change for SaaS providers is dizzying. SaaS vendors that view this only as a pricing problem will fail to become pivot-ready and thus struggle to keep pace. The only way for providers to get ready for the radically changed landscape ahead is to reground in the fundamentals of customer value, get the pricing game right, and build operational excellence that allows for experimentation and adaptation.