In a recent conversation with Lukas Egger on SAP’s Process Transformers podcast, I shared my perspective on how companies should approach pricing for AI-enhanced products. The conversation revealed critical insights about monetization strategy that could save SaaS companies millions in missed revenue opportunities.
Timestamped Overview
00:00 Hidden Costs in SaaS Models
06:02 Value Augmented, Not Revolutionary
08:09 Intercom’s AI Enhances Value
12:28 SaaS Pricing and Strategy Essentials
15:40 CFOs’ Concerns: Rising AI Costs
18:06 AI-Driven Tiered Pricing Strategy
21:02 Output-Based Ticket Resolution Pricing
25:00 Strategy: Monetization vs. Adoption
28:58 Defining Value in Customer Strategy
31:15 “AI’s Economic Impact Concerns”
Three Essential Approaches to AI Pricing
1. Be Intentional About Your Strategic Direction While Managing Real Costs
When approaching AI pricing, the first step is to be explicit about your strategic approach: are you optimizing for monetization or adoption? As I noted in the podcast:
“Do you want to maximize monetization, or do you want to maximize adoption? Have that as an explicit decision upfront.”
Many companies allow AI capabilities to penetrate their user base broadly before implementing strict monetization rules. This can be strategically sound if it’s a deliberate choice, not a default position.
This decision becomes even more critical given the unique cost structures that AI introduces. Unlike traditional SaaS, where marginal costs for additional customers approached zero, SaaS companies using AI now face an uncomfortable reality: bills from foundation model providers like OpenAI, Anthropic, and Google that scale with usage in previously unfamiliar ways.
Making an intentional strategic choice about monetization versus adoption helps frame how you’ll manage these variable costs internally while maintaining an appropriate customer-facing pricing structure.
2. Recognize That AI Doesn’t Require an Entirely New Playbook
Despite the rapid evolution of AI capabilities, the fundamentals of effective SaaS pricing haven’t changed. Many executives are scrambling to rethink their pricing models completely, but I disagree with this approach.
The reality? AI is simply an extension of a much longer trend of innovation and automation. Before jumping on the AI pricing revolution bandwagon, ask yourself: Does your AI create new value drivers, or does it enhance existing ones?
If your AI capabilities enhance an existing value driver (like helping a sales rep be more effective in a CRM), don’t fundamentally change your pricing model. Instead, consider:
- Adding AI capabilities as premium features in higher tiers
- Increasing per-user prices to reflect the additional value
- Maintaining the simplicity of your existing model
Intercom’s approach with AI summarization capabilities exemplifies this, enhancing their customer support agent experience while maintaining per-seat pricing.
3. The Underpants Principle: Create Outcome-Based Metrics, Not Cost Pass-Throughs
One of the most critical mistakes companies make when incorporating AI capabilities into their products is directly exposing their cost structure to customers. As I explained during the podcast:
“Don’t show customers your underpants. Just because your cost model has some underlying component – that is not relevant or interesting. Your customers don’t ask you how many lines of software are in your product. They don’t care where your developers are based. That’s your problem. That’s not their problem.”
Many SaaS companies are panicking in response to the variable costs of foundation models like OpenAI, Google Gemini, or Anthropic Claude. Suddenly, their margins are threatened by token-based pricing. However, passing this pricing structure directly to customers through token-based billing is misguided.
Your customers don’t want to understand what a token is any more than Uber riders want to see a line item for SMS fees on their receipt.
Outcome-based metrics make sense when AI replaces human work rather than augments it. This is where the underpants principle becomes particularly important – don’t charge based on tokens or API calls (your underlying costs), but on business outcomes your customers already measure and care about.
For example, Intercom’s Fin AI charges 99¢ per successfully resolved customer support ticket without human intervention. This works because:
- It aligns with clear, measurable outcomes
- Customers understand what they’re paying for
- The metric directly relates to how the customer is already measuring that department (ticket resolutions per agent per day)
The key difference? Tokens, API calls, and compute resources are your underpants – implementation details that shouldn’t be exposed to customers. Successfully resolved support tickets, completed transactions, or automated decisions are business metrics customers track and value.
Remember that Uber example: customers care about getting from point A to point B, not how many text messages Uber sends through Twilio during their ride. Similarly, your SaaS customers care about solving problems, not consuming tokens.
Value-Based Pricing Remains the North Star
Despite the complexities AI introduces, value-based pricing should remain your guiding principle. Understand what outcomes your customers try to achieve and how your AI capabilities help them get there more effectively. Price according to that value, not according to your cost structure.
Remember: in B2B SaaS, it’s not what you charge that determines your success. It’s who and how you charge.
Want to discuss how to optimize your SaaS pricing strategy? Reach out for a consultation at Product Tranquility or connect with me on LinkedIn.