The Death of Per-Seat Pricing: What It Means for Your SaaS P&L

Per-seat pricing isn’t dying quietly.

It’s being replaced in real time, and the financial implications for legacy SaaS companies are significant. Don’t ignore these market shifts if your product is a candidate for hybrid or outcome-based pricing.

Bloomberg is forecasting that subscription-based pricing drops from 60% to 30% of SaaS models over the next decade, while outcome-based pricing rises from 10% to 60%.¹ That’s not a small shift. That’s a full inversion of the model that defined SaaS finance for 20 years.

I’m not writing this to scare you. I’m writing it because most SaaS finance teams are not yet set up to handle the transition. I hate to say “pricing committees” but they should be meeting bi-weekly to discuss these shifts.

If your board and investors start to ask hard questions, you want to have your narrative in place.

What’s Actually Happening in the SaaS Market

The clearest signal isn’t in a research report. It’s in what enterprise buyers are actually doing.

SaaStr shared their own numbers: they went from 10+ human Salesforce seats to 2 human seats and 1 API seat, an 80% reduction in human users. But their Salesforce bill went up 83%, from $12,000 to $22,000 per year, because their 20+ AI agents use the platform roughly 100x more than humans ever did.²

That’s the new dynamic in a single data point:

Fewer seats. More usage. Higher spend.

But only for vendors who’ve already shifted to consumption pricing.

For vendors still on pure per-seat models, the math runs the other direction: fewer humans means fewer seats, fewer seats means less revenue, and there is no consumption upside to capture.

Per-seat pricing adoption has already fallen from 21% to 15% of SaaS companies in just twelve months, and companies clinging to pure per-seat models are experiencing churn rates 2.3x higher than those with hybrid or outcome-based billing.³

Software companies have already noticed. The PricingSaaS 500 Index tracked over 1,800 pricing changes across the top 500 SaaS and AI companies in 2025 alone — that’s 3.6 changes per company in a single year.⁴

A Bain & Company analysis of 30+ established SaaS vendors found 65% have already introduced hybrid pricing models layering AI usage charges on top of their base seat structure.⁵

The specific examples tell the story:

• Zendesk no longer charges purely by seat for its AI tier. It charges per resolved ticket — $1.50 to $2.00 per automated resolution. You pay when the outcome happens.⁶
• Intercom Fin charges $0.99 per resolution. The product earns revenue when it works.⁷
• Salesforce Agentforce launched at $2 per AI conversation, then expanded to multiple billing options after enterprise pushback: flex credits at $0.10 per action, per-user licensing at $125/month for unlimited internal use, and pure pay-as-you-go consumption.⁸
• Figma, HubSpot, Airtable, and Monday.com all adopted credit-based models in 2025.⁹ Credit model adoption surged 126% year-over-year across the top 500 SaaS companies.⁴

And it’s not just the AI-first vendors. The entire enterprise software stack is moving simultaneously:

• ServiceNow Now Assist is now tracking to $1.5 billion in ACV for 2026 — a massive upward revision from the prior $1 billion target. CEO Bill McDermott disclosed that 50% of net new business now comes from non-seat-based pricing, incorporating tokens and consumption metrics.¹⁰
• Workday reported AI-related ARR exceeding $400 million in Q4 FY26, with over $100 million in new AI ACV generated in Q4 alone, more than doubling year-over-year. Their Flex Credits consumption model now tracks API and agent usage across the entire platform.¹¹
• GitHub Copilot is moving to pure token-based billing on June 1, 2026. Your plan price becomes your monthly credit pool; no more flat seat fee with unlimited model access. A quick chat question and a multi-hour autonomous coding session will no longer cost the same amount.¹²

IDC data backs this up: 70% of software vendors are expected to move away from pure per-seat pricing by 2028.¹³

Gartner projects 40%+ of enterprise SaaS spend shifts to usage, agent, or outcome-based pricing by 2030.¹⁴

We’re not at the beginning of this shift. We’re in the middle of a long pricing shift.

How the Economics Change: A SaaS P&L Lens

The traditional SaaS P&L is built for predictability.

Subscription revenue comes in monthly or annually, smoothed and contracted. SaaS COGS are largely fixed in the short run with a linear path over time. That combination produced the 70%+ gross margins SaaS investors rewarded with premium multiples.

The new model breaks that structure in two places simultaneously.

Revenue becomes variable.

In an outcome- or consumption-based model, you don’t know exactly what you’ll bill a customer until they’ve consumed or until outcomes are delivered. Usage adoption curves replace contracted ARR as the forecasting input. That introduces a level of revenue variability we haven’t had to model before.

COGS become variable too.

AI inference costs are not fixed.

Every outcome delivered, every ticket resolved, every agent action, and every query processed consumes AI compute. And as you deliver more outcomes, your inference costs scale accordingly.

Unlike traditional SaaS COGS, which mostly scale with headcount and traditional hosted compute, AI COGS scale with product usage.

That’s double variability: variable revenue and variable costs moving in the same direction.

Your gross margin is no longer a stable band you can target. It becomes a ratio that fluctuates with usage intensity, model routing choices, and workflow complexity.

The 80% gross margin benchmark that SaaS finance teams have relied on for years doesn’t hold in an AI-native pricing model. ICONIQ Capital’s 2026 State of AI report, along with industry benchmarks, puts AI-feature gross margins in the 50% to 60% range, compared to 80% to 90% for traditional SaaS.¹⁵

That’s 30+ points of compression, and it’s not a rounding error.

It fundamentally changes how you structure your business.

What This Looks Like on a SaaS P&L: Before and After

Let’s make this concrete.

Take Acme SaaS, a $5M ARR company, 100% seat-based today, running a clean SaaS P&L.

Before: 100% Seat-Based

• Revenue: $5.0M, all subscription
• COGS: $1.0M, including hosting, support, customer success, and DevOps
• Gross Profit: $4.0M
• Gross Margin: 80%

Revenue is predictable, contracted, and smoothed monthly.

COGS are largely fixed. They step up slowly with more customers and more employees.

Now assume 30% of Acme’s revenue shifts to consumption or outcome-based billing as they embed AI agents into the product.

After: 70% Seat / 30% Consumption

• Seat Revenue: $3.5M, contracted and predictable
• AI Consumption Revenue: $1.5M, variable and fluctuating monthly with customer usage
• Traditional COGS: $0.9M, including hosting, support, customer success, and DevOps
• AI COGS: $0.75M, including inference costs, model routing, and agentic workflow compute
• Gross Profit: $3.35M
• Gross Margin: 67%

That’s a 13-point gross margin compression from a 30% revenue mix shift to a 50% margin AI product.

And critically, the COGS line is no longer stable. It moves with usage. A spike in customer adoption is good news, but it immediately hits your inference costs.

That creates P&L pressure. Your gross margin now fluctuates quarter to quarter in a way it never did under pure seat pricing.

This is the scenario every SaaS CFO needs to model before it happens, not after. And definitely before budget season.

COGS in the AI Era: The New Line Items

In my SaaS P&L framework, I teach COGS as the cost to deliver your product to a paying customer.

This includes support, customer success, services, and DevOps.

That definition still holds, but the composition is changing materially.

Traditional COGS

These are still there:

• Tech support
• Customer success
• Services
• DevOps

AI-Era COGS Additions

New AI-era COGS include:

• Inference costs or self-hosted compute
• API call costs to OpenAI, Anthropic, or your own model infrastructure
• Different queries routed to different models at different price points
• Agentic workflow compute
• Multi-step agents where each step is a separate inference cost

That last one deserves attention.

Multi-step agentic workflows can chain 10 to 100 inference calls for a single customer action. Each step has its own cost.

Tracking per customer or per month is no longer sufficient.

You need per-workflow cost tracking if you want to understand your true gross margin per outcome delivered. Time to meet with your CTO.

This is the new chart of accounts challenge. Just as I’ve always taught the importance of revenue stream clarity, you now need the same clarity on your AI cost structure.

If you can’t trace inference costs to the workflow that generated them, you’re flying blind on margin.

Start building that tracking infrastructure now, before you need it.

What SaaS Metrics Change: The Harder Question

This is where most finance teams are under-prepared.

NRR changes shape.

Traditional NRR expands by adding seats.

Usage-based NRR expands through consumption growth, which can be faster on the upside because more usage means more billing. But it is also more volatile on the downside.

A customer can cut usage in Q3, and your NRR contracts without a single churn event. The metric still matters. The behavior is different.

GRR gets modified.

With usage models, GRR is an unreliable metric.

You may want to drop the contraction input to the traditional GRR formula. Show a GRR metric that only counts churned dollars.

CAC Payback needs recalibration.

The standard formula uses gross margin.

In a model with variable inference costs, you need contribution margin by usage cohorts: heavy, medium, and light if you are charging a subscription.

Using blended gross margin overstates your payback efficiency if inference costs are significant or your user base skews toward heavy usage.

LTV is the biggest blind spot.

Traditional LTV calculations don’t account for variable cost-to-serve.

If inference costs scale with usage, your highest-usage customers may have materially lower LTV than their revenue suggests.

Traditional LTV formulas that use blended subscription gross margin, rather than contribution margin net of AI COGS, can meaningfully overstate true lifetime value in AI-heavy models.

New metrics worth tracking now:

• AI COGS Ratio: Inference and compute costs as a percentage of AI-driven revenue
• Inference Efficiency Ratio: AI revenue delivered per dollar of inference spend
• Compute-Adjusted LTV: LTV calculated using contribution margin, not gross margin

These aren’t hypothetical metrics for the future. If you’re already shipping AI features with any form of usage-based billing, you need these today.

In my Five Pillar Framework — retention, growth, efficiency, margin, and financial — AI economics now sits as a sixth pillar.

It’s not optional. It’s the lens that connects all the others in an AI-native business.

What To Do Now: Seven Steps for SaaS CFOs

You don’t need to have it all figured out. But you do need to start moving.

1. Audit your revenue mix.

What percentage of your ARR is currently seat-based or traditional SaaS? Usage-based? Hybrid?

If you don’t know the number, you’re not ready for the transition.

2. Start tracking inference costs separately.

Create a new line item in your chart of accounts for AI inference and compute costs.

Don’t bury it in hosting.

You can’t manage what you can’t see. See my AI COGS post for more details.

3. Model the P&L impact of a 30% seat-to-usage or outcome conversion.

If 30% of your current seat-based customers shifted to usage-based billing at 50% margins, what would happen to your revenue, COGS, and overall gross margin?

Run the scenario before the board asks for it.

4. Model the “double tax” risk on your vendor contracts.

As a buyer, your team may soon face both seat fees and consumption charges from the same vendor. ServiceNow, Salesforce, Workday, and GitHub are all layering credits and overages on top of existing seat structures.

Audit every major vendor contract up for renewal and model the worst-case overage scenario. The opaque overage structures in some of these hybrid models can surprise you at quarter-end or at renewal time.

5. Build per-customer cost tracking before you need it.

Per-workflow and per-customer cost attribution is table stakes for understanding your margin in a consumption or outcome pricing model.

6. Recalculate your SaaS metrics with variable costs.

CAC Payback, LTV, GRR, and NRR all need to be stress-tested with variable inference costs. If your unit economics look the same as two years ago, something’s wrong.

7. Present the transition plan proactively.

Don’t wait for investors or board members to flag the issue. Come to the next board meeting with a pricing model audit and scenario analysis. That’s what strategic finance leadership looks like in 2026.

The Pricing Bottom Line

Per-seat pricing isn’t dead overnight. I don’t believe all software models are meant for outcome-based pricing or agentic AI workflows.

However, the Bain research makes it clear that hybrid models are winning, not pure outcome-based billing.⁵

But the direction is set. Enterprise buyers are already making decisions that compress seat counts.

AI COGS are already eating into gross margins.

And the metrics we’ve relied on for a decade are already drifting out of alignment with the new economic reality.

The SaaS finance teams that will look smart at budget season are the ones building the infrastructure now:

• Cost tracking
• Scenario models
• Updated metric frameworks

Download: Pricing Model Audit Checklist

I’ve put together a one-page checklist you can use to audit your pricing model exposure and prepare for the transition.

It covers:

Revenue Mix Audit

What percentage of your ARR is seat-based vs. usage vs. hybrid? What does your renewal book look like by pricing model?

Vendor Double-Tax Calculator

For every major SaaS vendor contract up for renewal, model the scenario: Current seat cost + projected consumption/credit charges. What’s the worst-case overage?

P&L Scenario Model

What happens to your gross margin if 20%, 30%, or 50% of revenue shifts to consumption? What are your new COGS line items?

Metric Recalibration

Are your CAC Payback, LTV, GRR, and NRR calculations ready for AI? Which customers look different under the new math?

Five Questions for Your Next Board Meeting

The questions your board will ask about pricing model risk and the answers you should have ready.

Download the checklist below.

Go Deeper on SaaS Metrics and AI Economics

Check out my new AI finance course for software operators.

Footnotes

1. Bloomberg, as cited by Ishaan Verma, “AI Agents Force SaaS Overhaul: Pricing, Risk Dominate,” WhalesBook, April 11, 2026. Note: The original Bloomberg forecast is behind a paywall and could not be directly verified. WhalesBook attributes this forecast to Bloomberg using hedged language, and no direct Bloomberg source for these specific figures was located. Attribution should read: Bloomberg, as cited by WhalesBook, April 2026. Consider softening the body framing or leading with a directly verifiable statistic instead.
2. Jason Lemkin / SaaStr, “Why We Pay Salesforce 83% More Than Last Year. But Stopped Using Notion Entirely. The AI Agent Seat Problem Is Real,” SaaStr.com, April 27, 2026. Verified: 10+ human seats to 2 human seats plus 1 API seat; bill increased 83%, from $12K to $22K; 20+ AI agents using Salesforce roughly 100x more than humans.
3. Kyle Poyar, “The state of B2B monetization in 2025,” Growth Unhinged, January 2026. Primary source for both figures. Based on a survey of 240+ software and AI companies conducted April–May 2025.
4. Kyle Poyar and Rob Litterst, “What’s working in SaaS pricing right now,” Growth Unhinged, January 7, 2026. Source for the PricingSaaS 500 Index data: more than 1,800 pricing changes across the top 500 SaaS and AI companies in 2025, and credit-model adoption rising 126% year over year.
5. Bain & Company, “Per-Seat Software Pricing Isn’t Dead, but New Models Are Gaining Steam,” October 2025. Direct source for the 30+ vendor analysis. Bain found roughly 65% introduced a hybrid approach layering an AI meter on top of seat-based pricing.
6. Zendesk, “About automated resolutions for AI agents,” Zendesk Help Center. Verified: Pay-as-you-go pricing is $2.00 per automated resolution; committed usage pricing is $1.50 per automated resolution.
7. Intercom, “Fin AI Agent outcomes,” Intercom Help Center. Verified: Fin is priced at $0.99 per outcome.
8. Salesforce, “Salesforce Agentforce Pricing,” Salesforce.com, verified April 2026. Verified: Flex Credits at $500 per 100,000 credits, $0.10 per standard action, conversations model at $2 per conversation, and Agentforce Add-on at $125 per user per month for unlimited internal use.
9. Confirmed via primary sources: Figma and HubSpot credit models via Kyle Poyar / PricingSaaS, Growth Unhinged; Airtable AI credits via Airtable blog and support documentation; Monday.com AI Credits via Monday.com Help Center.
10. ServiceNow, “ServiceNow Reports First Quarter 2026 Financial Results,” and Q1 2026 earnings call, April 22, 2026; reiterated at ServiceNow Financial Analyst Day, May 4, 2026.
11. Workday Q4 FY2026 earnings call, February 24, 2026. Emerging-AI ARR exceeded $400 million; over $100 million in new AI ACV generated in Q4; 1.7 billion AI actions delivered across the platform in FY2026.
12. GitHub, “GitHub Copilot is moving to usage-based billing,” GitHub Blog, April 27, 2026. All Copilot plans transition to token-based billing on June 1, 2026.
13. IDC, “Is SaaS Dead? Rethinking the Future of Software in the Age of AI,” IDC Blog, December 2, 2025. IDC predicts that by 2028, pure seat-based pricing will be obsolete, with 70% of software vendors refactoring pricing strategies around new value metrics.
14. Gartner projection, widely cited: at least 40% of enterprise SaaS spend shifts to usage, agent, or outcome-based pricing by 2030. Note: The primary Gartner report is subscriber-only, and a specific public Gartner document for this exact figure could not be located.
15. ICONIQ Capital, “State of AI: Bi-Annual Snapshot,” 2026. AI-native product gross margins are projected to average roughly 52% in 2026, up from 41% in 2024, and below the 80%+ gross margins long associated with traditional SaaS.

Ben Murray

I have worked in finance and accounting for 25+ years.  I’ve been a SaaS CFO for 9+ years and began my career in the FP&A function.  I hold an active Tennessee CPA license and earned my undergraduate degree from the University of Colorado at Boulder and MBA from the University of Iowa.  I offer coaching, fractional CFO services, and SaaS finance courses.