Why B2B SaaS Pricing Is Broken (And What Is Changing)

The per-seat pricing model made sense when software was primarily about giving individual humans a tool. One person uses the tool, they get one license. The economics were clean.

AI-powered software is breaking this model. When a product can do the work of 10 people, per-seat pricing is either absurdly underpriced (from the vendor's perspective) or absurdly overpriced (from the buyer's perspective, who is paying for seats they are replacing with automation). The industry is in an awkward transition.

Why Per-Seat Pricing Worked for So Long

Per-seat pricing had virtues that made it dominant:

Predictability for buyers: You have 30 people who need access, you pay for 30 seats. Budget planning is simple.

Land-and-expand for sellers: Start with a department, prove value, expand to more seats. The growth model is built into the pricing structure.

Easy to administer: IT can count licenses. Renewals are easy. Overages are easy to bill.

These virtues still apply for many software categories. Zoom, Slack, Figma - tools where value is proportional to the number of users - per-seat pricing is still logical.

Where Per-Seat Pricing Breaks

Consider a legal research tool powered by AI that can prepare case research that previously took a paralegal 8 hours in 15 minutes. If a law firm has 10 paralegals and deploys this tool:

• Do they pay for 10 seats (even if 7 paralegals are now doing other work)?
• Do they pay per case researched (usage-based)?
• Do they pay for the time value delivered (outcomes-based)?

The vendor wants to capture the value they are delivering - 7.75 hours of paralegal time saved per case. The buyer wants pricing that reflects the inputs they are consuming, not the outputs they are getting.

This tension is playing out across every category where AI automates a meaningful portion of the work. ChatGPT Enterprise pricing, AI coding tools like Cursor editor, and AI customer support tools are all navigating this question.

The Pricing Models Emerging

Usage-based pricing: Pay per API call, per document processed, per query answered. This is the model most AI API providers use. It is fair in that you pay for what you use, but it creates unpredictable bills and requires sophisticated buyers to model costs before committing.

Outcomes-based pricing: Pay for results rather than inputs. A customer support AI charges per ticket resolved rather than per seat or per message. A contract review tool charges per contract rather than per user. This model is most aligned with value delivered but hardest to implement operationally.

Tiered capability pricing: Rather than seat tiers, capability tiers. The basic tier does document summarization. The professional tier adds workflow automation. The enterprise tier adds custom model fine-tuning. Pricing reflects capability rather than headcount.

Hybrid models: Most mature AI products are landing on hybrids - a base platform fee plus usage-based overages, or per-seat pricing for human users plus consumption pricing for AI-executed workflows.

What This Means for Buyers Right Now

The pricing structures you encounter for AI-powered products are genuinely still in flux. Several implications:

Negotiate aggressively on early contracts. Vendors in this space are still figuring out pricing. Early customers have leverage that later customers will not have.

Model usage carefully before committing. Usage-based pricing can produce surprising bills. Understand your actual consumption patterns before signing an annual contract with usage minimums.

Ask about pricing roadmaps. If you sign on at per-seat pricing and the vendor moves to usage-based pricing in 18 months, what happens to your contract? This is not a theoretical question.

Evaluate value created, not just cost. For AI tools specifically, a tool that costs $10,000/year and saves 500 hours of work is worth evaluating on different terms than a $10,000/year project management tool that saves 20 hours.

For detailed pricing comparisons across major software categories, see the best AI tools and best project management tools pages on ComparEdge.