In a recent presentation by Monetizely, the speaker provides a comprehensive framework for pricing and packaging AI SaaS products. This educational video breaks down the three critical decisions product teams must make when bringing new AI offerings to market: packaging structure, pricing metrics, and price points.
The Three Key Decisions in AI Product Pricing
When launching an AI product, you need to make three fundamental decisions:
- Packaging: What is your offer for your defined customer segment?
- Pricing Metric: On what basis will you charge customers (per user, per gigabyte, per token, etc.)?
- Price Point: For each package and pricing metric, what specific price will you set?
As the speaker explains: "These are the three decisions that we have to figure out how to make for our upcoming AI products."
Packaging Strategies: From Mass Market to Custom Enterprise
Packaging decisions exist on a spectrum based on market characteristics. On one end, there are mass-market offerings with simple tiers. As the speaker notes, "Netflix has one or two tiers at most… they are trying to reach the whole world and across countries they don't have a lot of tiers."
On the other end, enterprise-focused companies like Medallia often use à la carte pricing for large clients:
"For the larger deals they often have to package their product in more of an à la carte basis. They will talk to a financial service company, get an understanding of everything that is required, and then they will line item by line item custom scope these deals."
Why do enterprise vendors use this approach? The speaker explains: "Because the willingness to pay of a financial services company is very variable. It could be 2 million or 3 million. And the vendor is trying to maximize the return."
Adding New AI Features to Existing Products
For companies adding AI capabilities to existing products, the speaker offers a practical 2×2 rubric based on popularity and willingness to pay:
- High Willingness to Pay + High Popularity: Include basic functionality in all packages with premium versions in higher tiers
- Low Willingness to Pay + High Popularity: Include in all packages without extra charges
- High Willingness to Pay + Low Popularity: Offer as a premium add-on
- Low Willingness to Pay + Low Popularity: Include only in premium tiers
Real-world examples illustrate these approaches:
- Zendesk included AI-powered bots in all packages with usage limits
- Sentinel One offers their AI Security Assistant in advanced tiers while making their Agentic SOC Analyst an add-on
- PagerDuty created an expensive AI Ops add-on for customers with high willingness to pay
The speaker emphasizes that packaging decisions must be rational and based on customer preferences: "A more rational conversation needs to happen within the company as to what we are trying to do with a new feature, what is our monetization goal behind it, and what are our customer preferences and customer feedback."
Selecting the Right Pricing Metric for AI Products
The choice of pricing metric falls on a spectrum from fixed to variable. Fixed metrics include flat monthly fees or per-user pricing, while variable metrics might include per-token or usage-based pricing.
When selecting a metric, the speaker recommends evaluating potential options against seven key factors:
- Customer risk perception: Customers prefer variable pricing for riskier products
- Mental anchors: What pricing models are customers already accustomed to?
- Value alignment: Does the metric increase as customer value increases?
- Revenue potential: Can you effectively monetize increases in the chosen metric?
- Consumption patterns: How predictable is usage?
- Cost patterns: How do your costs scale with usage?
- Competitive actions: What metrics are competitors using?
The speaker provides a practical example comparing different metrics for a Zendesk-like chatbot:
"Dollar per resolution will be considered valuable by customers… It is definitely aligned to their success because the more cases are resolved, the more successful their customers are… but the definition of resolution is very fuzzy… I would have been very skeptical about this because it may align with customers but it may open up another can of worms."
For AI products specifically, usage-based metrics are increasingly common because of cost structures:
"A lot of companies are never going to use something like user-based pricing because… maybe somebody's going to use your product a lot and you'll start losing money. So you will be almost forced into using some sort of usage-based model."
Setting Price Points for AI Products
When determining price points, companies must consider:
- Market competitiveness
- Cost of goods sold (COGS)
- Product usage patterns
- Customer value perception
For AI products specifically, margins tend to be lower due to several factors:
"R&D costs are going to be higher because you have to develop some models of your own. There's going to be fine-tuning required… The cloud infrastructure costs are obviously going to be higher, data costs are higher, and there is an AI tax on the compliance side."
The speaker demonstrates this with a hypothetical calculation for a Zendesk-like product:
"For one customer of Zendesk that has 100 agents, if you used ChatGPT-4, it would cost… $55,000… which works out to $145 per agent per month, which is 40-50% of their per-agent cost. But if you use Deepseek, that comes down to $6 per agent per month."
This cost difference has profound implications for engineering decisions:
"You may try and experiment with something like 40, but as soon as you're successful, you need to switch over to some other model because you are bleeding money… That requires an understanding of economics and cost on the engineering side where previous CTOs have had no need to understand this."
Key Takeaways for AI Product Pricing
The speaker concludes with three essential points:
- Packaging is a function of your market: Consider customer segments and their willingness to pay
- Pricing metrics must balance customer and company success: "Don't blindly copy someone else"
- COGS for AI products are fundamentally different: Understand your unit economics deeply
As the AI product landscape continues to evolve, these pricing and packaging decisions will play a crucial role in determining which companies succeed. The speaker believes we're entering a new regime where "these products are built have a huge implication of how the next crop of AI companies are going to operate."
For SaaS executives navigating this new terrain, making informed, strategic decisions about pricing and packaging will be essential to capturing the full value of AI innovations while building sustainable business models.