How AI Is Forcing Companies to Rethink Value Creation from First Principles (Insights from Kyle Poyar on The Meg & Amy Show)

There's no perfect pricing model for any business. But with AI, there are suddenly a lot more wrong ones.

"It's really about being aware of the trade-offs you're making," Kyle Poyar - creator of Growth Unhinged - told us during our recent conversation. "What's right for your product? What allows you to tell the story around what you're building, who you're building it for, how it helps customers be more successful."

The challenge is that AI is breaking the fundamental assumptions underlying every traditional software pricing model. The story that worked for seats, users, and feature tiers? It doesn't work when a single AI interaction might deliver massive value or fail completely, while consuming unpredictable and escalating compute costs.

After 15 years helping software companies scale from $1M to $100M+ ARR, Kyle sees this moment as more than just a pricing challenge—it's a complete rethinking of how value gets created and captured in software.

For decades, software pricing followed predictable patterns: on-premise licenses with high upfront costs, then subscriptions that democratized access, then product-led growth (the term Kyle coined at OpenView) that required proving ongoing value. Each evolution had familiar pricing primitives—seats, users, feature tiers—that aligned costs with value delivery.

AI shatters this alignment. Traditional software scaled predictably: more users meant more value delivered and proportional infrastructure costs. But AI can deliver massive value to a single user while consuming wildly variable compute resources. One AI conversation might resolve a complex issue that would have required multiple human interactions—or it might fail completely, burning expensive tokens in the process.

The flat fee trap: Initially, many companies tried bundling AI features into existing subscriptions, betting that AI costs would plummet. "There were a lot of people that thought AI costs were going to drop 10x every year," Kyle explained. "Why worry about AI monetization right now? It's fine if we have negative margins."

But that's not what's happening. "People are now seeing that that's not what's playing out. The LLMs are not necessarily getting that much cheaper. Your customers want access to the best technology. They don't want yesterday's newspaper."

In fact, it's the opposite: "AI costs aren't dropping 90%. They're increasing by orders of magnitude potentially" as newer models consume far more tokens while delivering enhanced capabilities.

The core problem: AI breaks the fundamental assumption that value and cost scale together predictably.

Kyle identified a fascinating trend that perfectly illustrates the complexity of AI pricing: "In the last three months, everyone seems to have decided we should launch some sort of credit model."

The pattern started with Salesforce, which initially charged per conversation for AI features, then quickly pivoted to a "flex credit model" where credits are consumed based on the complexity of AI actions. "If it was doing more to fully resolve a conversation, that was going to be more expensive than if you just enriched a contact or updated a CRM record."

This credit approach has now spread across the industry. But as Amy shared from her recent experience at SAP, the execution is revealing critical challenges: different buying centers (HR, finance, IT) sharing credit pools without visibility into each other's usage, and unpredictable costs that one customer compared to "when their teenager had a usage-based cell phone and they got like a $900 phone bill."

Kyle's analogy perfectly captured the customer experience: "It's like you're having a kid's birthday party. You let the kids go to the arcade and all of a sudden you have a bill for hundreds or thousands of dollars based on all the games they played."

When we asked Kyle where he believes pricing models will be by 2030, his answer was clear: true outcome-based pricing enabled by AI capabilities.

"The model I expect to see a lot more of in 2030 is being really, really clear on the outcomes that the customer is trying to get, and proving that you're able to deliver at those outcomes at a rate that competitors can't match."

He pointed to Intercom as an early example: They said, "For our AI product, we're only going to charge you $0.99 per conversation that AI resolves. We're going to define a resolution as X, Y, Z factors. We're also going to let you set cost thresholds so that you have control over budget."

What makes this powerful isn't just outcome alignment—it's the complete system: clear success definitions, customer budget controls, performance transparency, and tools to improve AI success rates over time.

"We need to find a way to compete that's tied to being able to deliver at a higher quality than competitors, or deliver at similar quality but at lower cost," Kyle explained. "The best way to pitch that message is to be really clear on the outcomes."

But as Kyle acknowledged, "It is really hard to make all of the changes necessary to get there."

This connects to themes Meg and I have been exploring about the fundamental challenge facing companies today. Most organizations are approaching AI as an efficiency play rather than a transformation opportunity.

The reality is that we're at a moment that requires turning everything upside down. As we explored in "The Death of SaaS As We Know It," AI is "unlocking entirely new markets and capabilities that didn't exist before." When business models can generate 5x the revenue per employee, the entire foundation needs to be rebuilt.

But here's where most companies get stuck: they're trying to retrofit AI onto business models designed for human leverage. As entrepreneur Usman Sheikh told us, "You're trying to augment people into workflows which need rethinking."

This is why Kyle's insight about "telling the story around what you're building" becomes so critical. When Intercom charges $0.99 per resolved conversation, their pricing tells a clear story: we deliver outcomes, not access. But when companies charge flat fees for AI features, the story breaks down.

This is why pricing matters so much. It's not just about revenue—it's a proxy for your entire business model. When AI breaks traditional pricing assumptions, it signals that everything else needs to change too.

The companies that recognize this as a transformation opportunity will define entire industries. Those who treat it as a pricing problem will find themselves competing against business models they never saw coming.

Kyle's most important insight was about fundamentally changing how companies think about pricing: "Many companies still have a technology-driven product organization where they focus on what they're going to build, and then they go, 'We're about to launch this. How much are we charging for this again?'"

The old approach—build first, price later—is broken in the AI economy. Instead, Kyle advocates for having pricing conversations with customers before you build, not after. But these aren't traditional pricing discussions.

The conversations companies need to be having:

Instead of asking "What would you pay for this?" Kyle recommends understanding the buying psychology: "What is your approach today? What's working about that? What's not working about it? How many hours does your team spend doing this? How hard is it to hire for this role? If you could adopt technology that would solve this for you, how interested would you be in buying that? Would this come out of technology or headcount budget?"

The education companies need to provide:

Kyle pointed to Intercom's approach of educating customers about how to buy AI: "Here's the typical cost per conversation when people resolve conversations. Here's the cost if you use AI plus people. Here's the cost if you use AI plus people and you build it yourself. And here's how you can evaluate the pros and cons between those different decision points."

The way companies need to think:

The goal isn't finding the "perfect" pricing model—it's choosing approaches that align with how you create and deliver value. As Kyle noted, successful companies "have a real thoughtful plan for how you're going to adapt" as markets and technology evolve.

This connects to broader themes Meg and I have been exploring about how companies approach AI transformation. As we discussed in our recent article on business model transformation, most organizations are trying to "retrofit AI onto business models designed for human leverage" rather than rethinking value creation from first principles. Muscle memory is hard to overcome.

What pricing model are you using or contemplating for AI features? How does it match with your value creation narrative? Share your thoughts in the comments.

Prepared by Amy Wilson, former tech executive and current product strategy advisor. For more insights on growth, leadership and the future of work, subscribe to The Meg and Amy Show.