How Intercom Built Its Outcome-Based Pricing Model for AI

Intercom’s pricing transformation began with a stark realization: a company known for innovation had lost customer trust in one of its most essential business components. In this episode of Thank You for Vibe Pricing, Aisling O’Reilly, Head of Pricing at Intercom, joins Chargebee’s Ariela Bitran and Vinay Seshadri to break down the decisions that reshaped Intercom’s model—and established the foundation for pricing Fin AI on outcomes instead of the usual activity-based levers like seats or agent usage that have dominated the customer service category for years.

The conversation offers an operator’s view into how a major SaaS company changed course: why Intercom needed to simplify its pricing, how it designed Fin’s outcome-based model, and what it took internally to support that shift.

A turning point in Intercom’s pricing strategy

Intercom reached a point where pricing was undermining customer trust. Years of strategy shifts and repeated pricing adjustments had created unnecessary complexity, resulting in unpredictable bills and customer frustration. Pricing had become both a brand issue and a structural blocker, creating complexity that eroded trust, drove churn, and made it harder for the business to operate and grow.

O’Reilly is direct about the urgency:

Rather than nibbling around the edges, leadership reset the entire pricing foundation. They ran scenarios showing as much as $50 million in ARR exposure, but made the call anyway. The eventual impact was closer to $15 million, proof that transparent, predictable pricing not only rebuilt trust but also unlocked greater customer spend.

The eighteen-month migration reset the foundation Intercom needed before introducing more ambitious monetization changes.

Designing Fin’s outcome-based pricing model

Fin’s launch highlighted a key tension: seat-based pricing does not align cleanly with an AI system that resolves conversations autonomously. Seats historically represented human capacity; for AI, they obscured value.

Intercom explored the industry’s default—usage or conversation-based pricing—through customer research, not a live rollout. The feedback was clear: buyers were uneasy about paying per interaction while still carrying the full cost of their human teams.

“At the time, AI agents felt so new and unknown. There was this sense of: this might not work, and if I’m paying per usage, I still have to pay all my human agents.”

This pushed the team toward a single outcome everyone could trust and measure: a conversation fully resolved by Fin with no human intervention. That definition became the basis of Fin’s pricing. It is simple to explain, straightforward to forecast, and directly aligned with ROI.

Related reading: Pricing AI agents: What does ‘value-based pricing’ really mean for AI?

Preparing to price outcomes

Charging for outcomes requires more than a pricing model; it requires instrumentation. As O’Reilly puts it, many AI products deliver value, but teams “don’t always have visibility on the value they’re providing in a way that they can reliably and transparently charge for it.” Customer support was different.

Intercom had been tracking conversation resolution long before Fin existed, giving the team a clear definition of success and the ability to validate that definition with customers. Its billing infrastructure could also meter nontraditional events, making per-resolution pricing operationally viable.

This readiness across definitions, metrics, and billing systems was one of the quiet enablers behind the move to outcomes.

Solving the predictability problem

Outcome-based pricing strengthens value alignment, but it can make forecasting more challenging. Support leaders often know how many human agents they need; they do not always know how many resolutions an AI system will handle early on. Intercom introduced several mechanisms to reduce that uncertainty:

• Free trials. Trials provide customers with an early indication of resolution rates and help them estimate expected spend.
• Annual resolution buckets. Instead of monthly quotas, customers draw from a yearly pool of resolutions that reflects the seasonal nature of support traffic.
• Non-penalizing overages. Where many vendors charge more once customers exceed contracted usage, Intercom applies the same contracted discount to all usage, including overage. Customers can also choose pay-as-you-go.

These decisions reduced CFO friction and enabled Intercom to maintain a simple model without compromising predictability, an approach they also detailed in Chargebee’s State of Recurring Revenue and Monetization report.

In that spotlight, CTO Darragh Curran captured the core tension of outcome-based pricing: “A surprising challenge has been predictability getting in the way of usage.” He described how the team addressed this through annual resolution buckets, non-punitive overages, pay-as-you-go options, and flexible paths to adoption.

Why AI is reshaping the product management role

As Intercom deepened its investment in AI, cost structure became inseparable from product design. Traditional SaaS products often operate with high margins. AI features introduce infrastructure and token costs that scale with usage.

Commercial awareness is now part of product management at Intercom. Pricing and product teams collaborate early to understand the cost dynamics of new capabilities and the value they are intended to create.

A single example illustrates the shift. Early versions of Intercom’s AI co-pilot included suggested prompts meant to help agents. These prompts consumed nearly half of the product’s operating cost, yet usage was under 1 percent.

This level of scrutiny is becoming standard for AI-driven products and essential for sustainable pricing models.

Where outcome-based pricing may evolve next

Fin’s resolution-based model captures a clear outcome, but it doesn’t capture every dimension of value. Fin regularly clarifies customer questions, gathers information, and routes conversations more efficiently. This reduces effort for human agents but doesn’t count as a resolution.

Intercom also provides advanced insights and testing environments for Fin, capabilities that carry meaningful cost but are not priced separately today.

O’Reilly acknowledges that the model will need to evolve as Fin becomes more capable:

Assisted outcomes and AI-evaluated success metrics represent possible future directions, but simplicity remains the guiding principle.

Why AI credit-based pricing is limited today

Credit-based pricing is becoming increasingly common among AI vendors, but O’Reilly draws a clear line between what credit models should be—value-based pools customers can allocate across products, outcomes, or usage—and what the industry often offers today: cost-plus structures tied to token consumption.

In Intercom’s customer research, buyers often described these cost-plus systems as difficult to forecast and misaligned with value, in part because they shift most of the cost risk to the customer.

Intercom’s model takes the opposite stance. As O’Reilly explains, “Our costs scale with how often Fin is in a conversation, not how often it resolves it. So we’re very incentivized, and we’ve taken on all of the LLM cost risk by having an outcome-based model.”

She expects cost-plus credit systems to be short-lived, though she sees long-term potential in value-based credits tied to outcomes rather than infrastructure usage.

Listen to the full conversation

For pricing, product, and finance leaders evaluating how to monetize AI systems, Intercom’s experience provides a practical perspective on the technical, operational, and organizational elements necessary to support outcome-based pricing.

Even more insights

Discover lessons from the field on using pricing to drive adoption, build sticky habits, and create an unfair advantage in AI markets.

Get the Pricing for AI Distribution Playbook →