AI Is Killing SaaS Pricing: Lessons From 10k+ Customers | Prittam Bagani, VP of PM (Chargebee)
AI isn’t just transforming how we build products; it’s rewriting the rules of how we price and sell them.
TL;DR
In this episode of LaunchPod, we sit down with Prittam Bagani, VP of Product Management at Chargebee, to unpack how 10,000+ SaaS companies are navigating the biggest shift in pricing strategy since the rise of cloud.
Here’s what we cover:
How AI is forcing companies to rethink pricing - and why usage-, outcome-, and agent-based models are the future (02:25)
Why OpenAI made over 100 pricing page changes in a single year and what you can learn about agile pricing experimentation from it (06:28)
And how to build the necessary technical and product infrastructure to test pricing models in real-time across millions of users (18:37)
1. AI is forcing a full-stack rethink of pricing
As AI tools evolve from simple workflow accelerators to full-on agentic replacements, pricing strategy has become a competitive weapon. Prittam outlines how AI-native companies are shifting away from legacy pricing models in favor of usage-, outcome-, and agent-based models, each designed to better match how customers extract value.
He breaks it down like this:
Usage-based: Common with tools built on LLM APIs. Think token counts, content generated, or API calls (e.g., OpenAI, Cursor)
Outcome-based: Products like Intercom are experimenting with charging per resolution, not conversation volume
Agent-based: Pricing around a human-equivalent function, like a sales assistant or support agent, opens the door to new budgets and value anchors
“Because of AI, people are starting to think about agent-based pricing models... If you're pricing it as an agentic model, you’re likely to tap into the budget that is 10x bigger than your typical software budget.”
2. What OpenAI’s 128 pricing page changes can teach you about agility
AI-native products move fast, and pricing strategies must keep pace. OpenAI made 128 changes to its pricing page in just one year, a clear signal that static, annual pricing reviews are no longer viable.
Here’s what product teams can learn:
Treat pricing like a sprint, not a quarterly task. As products evolve rapidly, pricing needs to be tested and adjusted just as frequently
Run controlled experiments with new pricing models before rolling them out broadly. Start with small cohorts to learn what works
Communicate changes clearly and in advance. Many pricing failures (like Cursor’s $20 → $500 backlash) come down to poor expectation setting, not the pricing itself
Use transparent usage data to help customers anticipate their bills. Forecasts, alerts, and usage dashboards reduce friction and surprise
Hybrid models help balance agility and predictability. Combining flat subscriptions with usage-based overages gives both finance and users confidence
Agile pricing isn’t just about speed; it’s about removing risk for the buyer, aligning closer to value, and staying ahead of the market.
3. What it takes to build real-time pricing infrastructure at scale
Experimenting with pricing in a spreadsheet is easy. Doing it across 50 million users in real time is not.
Prittam walks through what it takes to actually build usage-based or outcome-based pricing that doesn’t break customer experience or finance.
“There are 50 million subscribers. They have the ability to, in real time, track your usage... and show a paywall that says, ‘Hey, you just hit the limit. Either upgrade or buy additional credits.’”
To get there, he says companies need to:
Instrument deep usage tracking: Capture every relevant event to feed value metrics
Define and align on value metrics: E.g., resolutions, content generated, tokens consumed
Layer in monetization logic: Use real-time usage thresholds to trigger gating, alerts, or upgrade prompts
Maintain financial predictability: Combine usage-based overages with flat-fee subscriptions (hybrid models)
He emphasizes that this is a product and technical problem, not just a business decision. Teams often underestimate the effort, burn engineering cycles, and still fall short:
“Don’t underestimate the technical complexity... It gets really complicated really quickly.”
For example, Chargebee has customers that hit scale so fast that billing, not product, became the #1 growth bottleneck. In some cases, sales-led and product-led motions coexist, and the billing system has to handle both flexibly.
Links
Chapters
02:25 Pricing Models in the AI Age
06:28 Challenges and Best Practices in Outcome-Based Pricing
18:37 Technical Complexities and Real-Time Pricing Adjustments
22:01 The Power of Timely Upgrades
23:07 Hybrid Pricing Models: The Future
24:57 Complexities in Sales-Led Growth
28:10 Unit Economics and Usage-Based Pricing
30:57 The Strategic Role of Product Teams in Pricing
33:58 The Competitive Edge of Agile Pricing
What does LogRocket do?
LogRocket's Galileo AI watches user sessions for you and surfaces the technical and usability issues holding back your web and mobile apps. Understand where your users are struggling by trying it for free at LogRocket.com.
Thanks for sharing! Love it when people share pricing lessons.
If I may challenge the guest a bit on his statement: "If you're pricing it as an agentic model, you’re likely to tap into the budget that is 10x bigger than your typical software budget.”
Sounds great, but this seems based on absolutely nothing. No data to back this up at all. No rationale behind it. Using the right pricing metrics can definitely increase the ACV by 2–3x, but not 10x.