How SaaS Pricing Is Changing With AI

For more than two decades, SaaS pricing followed a predictable formula: charge companies per seat, per month, and grow revenue as accounts grow. It worked beautifully in the cloud era. Salesforce, HubSpot, Slack and thousands of others scaled into billion-dollar companies using variations of the same model, but that reliable formula is breaking down with AI.

When AI enters a product, the economics change. Costs fluctuate with compute usage. A single user can generate thousands of AI tasks. And the value delivered may no longer depend on how many employees log in. Instead, the value might come from what the software accomplishes.

It's driving one of the most dramatic shifts in SaaS pricing in decades.

As software becomes the infrastructure layer for AI agents and automated workflows, the SaaS model is evolving to reflect that reality.

AI is Breaking the SaaS Model

At the center of the shift is a simple but powerful force: AI changes the cost structure of software.

Traditional SaaS had relatively predictable costs. Once a product was built, adding more users rarely increased infrastructure costs dramatically. AI products are different — each prompt, inference or workflow triggers compute usage, which creates real costs for software companies. When one user might generate 10 requests per day and another might generate 10,000, it’s difficult to sustain the traditional SaaS pricing model.

Stripe explains: “The old software-as-a-service (SaaS) model with flat subscriptions and per-seat licenses often breaks for AI, because the cost of serving each user shifts with computing demand.”

At the same time, AI dramatically changes how customers experience value. One employee using AI might produce the output of five employees without AI. If pricing is tied only to seats, software vendors may capture only a fraction of the value they create.

If SaaS is going to survive, it has to re-imagine pricing systems to balance three forces:

1. Cost to serve, addressing the real marginal costs of AI inference.

2. Customer value based on productivity gains from AI.

3. Pricing simplicity that maximizes adoption.

---

Extend your runway and keep pace with AI.

Lighter Capital has provided over 1,000 rounds of financing to tech startups, totaling over $500 million. Apply in minutes to get up to $10 million in non-dilutive growth capital for your startup — no pitch decks, presentations, or business plans required.

GET FUNDING

---

When Seats No Longer Scale

AI is pushing companies to stop charging for software and start charging for value delivered. Investor firm Bessemer Venture Partners describes the shift succinctly: “Companies are no longer just selling access; they’re selling outcomes.”

Leading software companies are experimenting with usage-based pricing, bundling in AI, and outcome-based pricing to better reflect the value the software actually delivers and protect their margins.

Usage-Based Pricing (UBP)

One of the biggest SaaS pricing changes is the rise of usage-based pricing (UBP). In a usage-based model, customers pay according to consumption. That could include:

• API calls

• AI prompts

• data processed

• messages sent

• workflows completed

• AI resolutions delivered

This model aligns well with AI because both costs and value scale with activity.

When AI product pricing tracks usage, rather than user access, there's a tighter relationship between the price customers pay and the value they receive.

Real-world example: Snowflake

Snowflake charges customers based on three factors: compute usage (data processing time), storage used, and data transfer. Compute usage is measured in credits, which increase as queries run longer or process more data. This aligns pricing with actual platform utilization.

Bundling AI Into SaaS Subscriptions

Not every company wants to meter AI usage. Some companies are taking a different approach: bundling AI into core subscriptions.

Bundling in AI reduces friction and accelerates adoption. Research from SBI found that when AI is sold as a separate add-on only 20% of net new customers purchase it, and only 38% of those buyers actually use it. That means only about 8% of all new customers end up using the AI feature the cost is not included.

Real-world example: Google Workspace

Google Workspace provides an example. Starting in 2025, Google began including many Gemini AI capabilities directly within Workspace plans rather than offering them solely as add-ons.

Outcome-Based Pricing

Outcome-based pricing is likely the future, where customers pay based on measurable results. This model is just starting to build steam. SBI reports that 0.6% of SaaS companies currently use outcome-based pricing.

Why so low? Because outcomes can be difficult to measure. AI agents make this more feasible by performing discrete tasks that can be measured, but you still have to track performance accurately, prove attribution, and manage your financial risk if results (and thus your revenue) fluctuates.

Ultimately, outcome-based pricing only makes sense when the outcome is clear and easy to measure. Outcomes might include:

• Revenue generated

• Deals closed

• Support tickets resolved

• Engineering tasks completed

Real-world example: Intercom

Intercom provides a good example of AI-driven pricing evolution. Its AI support agent, Fin, is priced at $0.99 per resolution. Instead of charging only for seats, Intercom charges for the work the AI performs: resolving a support request.

Hybrid Pricing Models Are Meeting the Moment

Pure usage-based pricing makes buyers anxious, but a simple all-in-one subscription puts your margins at risk. How do you support adoption without collapsing your unit economics? Offer a mixed pricing framework that aligns with the customer’s perceived value and your variable costs.

According to SBI, companies aren’t abandoning their legacy subscription plans entirely — 80% of software companies still use some sort of seat-based pricing.

3 Hybrid SaaS Pricing Models

1. Primary or secondary seat-based component: Of the 80% still using seat-based pricing, nearly 90% combine that with another value metric. Just 8% are pricing based on seats, only.

2. AI usage in a flat fee plan: 37% of software companies have adopted this pricing structure, making it the most common hybrid pricing model.

3. Flat fee plans with usage-based AI add-ons: 23% of SaaS companies are offering this combination of pricing.

This hybrid approach provides predictability and flexibility, allowing pricing to scale with activity.

Real-world example: Salesforce

Salesforce’s AI platform Agentforce illustrates how hybrid pricing works. The company offers both traditional per-user licenses and consumption-based pricing via Flex Credits or conversation usage. This structure allows its customers to adopt AI incrementally while paying for additional AI activity as needed.

Real-world example: Microsoft Copilot

Microsoft also blends licensing and consumption. Organizations can use certain Copilot features with pay-as-you-go billing, allowing them to test AI usage patterns before committing to full licensing.

Choose-Your-Own Pricing

Instead of forcing customers into one pricing model, some companies are allowing customers to choose a combination of pricing frameworks from options, giving them control and flexibility. Similar to other hybrid pricing models, these options might include:

• Seat-based plans

• Usage-based plans

• Consumption credits

• Bundled AI features

• Custom enterprise contracts

This “choose-your-own-pricing” approach gives customers maximum control and helps them adapt to widely varying AI usage patterns. It also addresses the reality that no two customers use AI the same way.

---

Explore SaaS Pricing Trends

Click to expand the infographic.

Source: SBI Growth Advisory. (2025). 2025 state of SaaS pricing report | Part 1: 10 insights for building market-leading pricing.

---

What Pricing Strategy is Right for Your SaaS?

Pricing strategies across the SaaS ecosystem vary widely. We see a couple factors steering businesses toward different pricing models as they account for AI.

AI-Native vs. AI-Enhanced SaaS

AI-native SaaS companies build products around AI from the beginning. Examples include AI coding tools, AI support agents, and generative design tools. These companies are more likely to adopt:

• Usage-based pricing,

• Credit systems; and

• Outcome-based pricing.

AI-enhanced SaaS — traditional SaaS companies adding AI features — often keep existing pricing structures. As a result we tend to see them:

• Bundling AI into higher pricing tiers,

• Charging for AI add-ons; or

• Introducing usage limits.

Vertical vs. Horizontal SaaS

Horizontal SaaS solutions serve many industries, for example, productivity software or marketing and sales platforms. These companies tend to favor usage-based pricing because AI workflows and use cases vary widely from customer to customer.

Vertical SaaS focuses on a specific industry such as: healthcare, construction, legal, or logistics. Because vertical SaaS platforms often sit deeper within workflows, they have more opportunities for outcome-based pricing tied to operational metrics.

Vertical SaaS may be quicker to transform their pricing models with AI that can complete a transaction, such as processing an insurance claim or medical record.

What Should Startup Founders Do Now?

For founders building AI-driven SaaS products, pricing strategy is quickly becoming as important as product design. Here are several practical steps to take to make sure your pricing can support a sustainable business.

1. Start with value metrics

Ask a simple question: What work does the AI actually perform? That work should influence pricing. Examples might include:

• Documents analyzed

• Workflows automated

• Tasks completed

• Leads generated

2. Start tracking usage early

Even if you launch with subscription pricing, start measuring:

• Prompt volume

• AI task completion

• Compute costs

• Workflow frequency

These metrics will help guide future pricing experiments.

3. Consider hybrid pricing

Hybrid models provide the best balance at this point in time. A hybrid pricing structure commonly includes a base subscription with either AI usage included or additional consumption pricing.

4. Don’t over-complicate it

AI pricing can become confusing quickly. Resist the temptation to create elaborate credit systems unless they clearly benefit customers.

5. Align pricing with outcomes over time

While outcome-based pricing may be difficult early on, it’s never too soon to start thinking about how to price around results, eventually.

SaaS + AI: The Big Picture on Pricing

SaaS is evolving into something far more powerful. The platforms that once delivered software through the cloud are now becoming the operating systems for AI-driven work. As AI agents handle more tasks once performed by humans, pricing will inevitably move closer to the value those systems produce.

So while the per-seat license may slowly fade, the underlying SaaS platforms are becoming even more instrumental.

In the Platform + Agent era, software pricing will increasingly track to productivity, automation and outcomes. And for founders building the next generation of AI-powered products, transforming pricing is proving to be both a challenge and a massive opportunity.