AI Doesn’t Charge by the Hour. Should You?

As a product leader with two decades in this field, I've seen firsthand how crucial pricing and packaging are to a product's success. It's something we, as product leaders, need to own, collaborating closely with other teams but ultimately driving the strategy.

Now, with AI becoming central to existing products and powering entirely new ones, the big question on everyone's mind is: 

This isn't just an academic question: 

❌ Getting it wrong can stifle adoption, frustrate customers, and hurt our business. 

✅ Getting it right can unlock tremendous value. 

As part of the GenAI.Works community, where we're all exploring the impact of AI on our roles, this topic is vital.

The Elephant in the Room: User-Based Pricing Doesn't Fit AI

Let's start with the most common model in SaaS today: user-based pricing (charging per seat or per user). I believe this model is fundamentally broken for the age of AI.

Why?

👉Traditional software usually helps employees do their jobs better or faster.

👉 AI often replaces parts of a job, or potentially entire roles.

Think about it. If AI handles 50% of the tasks previously done by 10 employees, you might only need 5 employees using the tool. Your potential user count shrinks, even as the value delivered to the company (like reduced staffing needs) increases significantly. The user-based model fails to capture this value.

Companies heavily reliant on user-based pricing face a tough challenge. They risk falling into the "innovator's dilemma," where newer companies use both cutting-edge AI and disruptive business models better suited to AI's value. Furthermore, moving customers away from familiar per-user pricing requires significant customer education and change management—a challenge in itself.

Expect big shifts ahead.

Exploring AI Pricing Models

So, if charging per user is out, what are the alternatives?

Cost-Plus Pricing

• What it is: Calculating underlying costs (compute power, API calls) and adding a margin.

• Why it's common: Simple to calculate and explain. Protects vendor margins.

• The challenge: Focuses on vendor costs, not customer value. It doesn't automatically prove the product delivers enough value compared to the alternative, which is often just sticking with the current manual process (or "doing nothing").

Activity-Based Pricing

• What it is: Charging based on specific, measurable actions the AI performs (e.g., per call made by an AI Sales Development Representative).

• Why it makes sense: Aligns with existing business metrics (like call volume).

• The challenge: Can lead to price wars (or "race to the bottom") for generic activities.

Workflow-Based Pricing:

• What it is: Charging for completing a more complex process (e.g., the entire SDR workflow from lead intake to researched call).

• Why it's better: Represents a more complete unit of value, harder for competitors to commoditize.

• The challenge: Requires clearly defining and agreeing with the customer exactly what steps constitute the completed "workflow."

Outcome-Based Pricing:

• What it is: Charging based on the successful result (e.g., per qualified meeting booked and attended by the AI SDR). Often the ideal model.

• Why it's powerful: Perfectly aligns vendor success with customer goals.

• The challenge: Defining and agreeing upon what constitutes a "successful outcome" can be complex. Furthermore, measuring these outcomes accurately and reliably often involves significant technical hurdles.

Replacement Fee Pricing:

• What it is: Pricing the AI relative to the cost of the human labor it replaces (e.g., AI SDR at $20k/year vs. human SDR at $60k/year).

• Why it's compelling: Direct cost-savings value proposition.

• The challenge: Global labor rates vary. Ensuring profitability when pegged to potentially falling labor costs is tricky.

Hybrid Models as a Path Forward

It's also important to note that these models aren't always mutually exclusive. Hybrid approaches are common and often practical. This could mean a small base platform fee plus charges per activity or workflow, or perhaps tiered packages that include certain volumes. Hybrids can be a good way to transition customers or balance predictability with value capture.

Bringing it Back to Business Health: Measurement, Margins, and Transparency

This brings us full circle. While value-based models are attractive, the practicalities matter. Ultimately, any pricing model you choose must support a healthy business. You need:

• Profitability: The price must cover all costs and generate the required margin.

• Measurability: You need robust technical systems to accurately track whatever you're charging for (usage, activities, workflows, outcomes). This technical implementation of tracking itself can be a significant hurdle.

• Transparency & Trust: Customers must clearly understand how they are charged and trust that the measurement is fair and reflects the value they receive. This reinforces the need for good communication and education.

Pricing AI is one of the most critical strategic challenges for product managers today. We must think beyond outdated models and embrace approaches that truly reflect the unique, often task-automating or outcome-driving, value AI delivers.

What pricing models are you exploring or implementing for your AI products? Share your thoughts in the comments!

About the Author

Filip Szymanski has over 20 years of experience as a product leader and is passionate about leveraging AI to power a new generation of product managers. Follow my journey at productpath.ai or add me on LinkedIn.

🚀 Boost your business with us—advertise where 10M+ AI leaders engage

🌟 Sign up for the first AI Hub in the world.

📲 Our Socials