Chargebee

Earlier this week, Intercom rebranded to Fin, and it perfectly encapsulates the shift we’re seeing in every corner of the industry: companies are moving from selling access to software, to selling outcomes driven by AI. Perhaps no function needs to change more as a result than finance. At Beelieve ’26 in San Francisco, Fin’s VP of FP&A Tommy Bettles sat down with Chargebee COO Jeff Sant and shared what that transformation actually looks like inside the business. Fin charges per successful resolution from its AI agent. That means revenue is no longer forecasted from contracts or seats alone. Finance now has to model an entire chain of product behavior: → How many customer conversations happen → How often the AI gets involved → How often it successfully resolves the issue → What each interaction costs to run → How much of that activity becomes billable revenue As Tommy put it: “It’s a whole new day” for the finance org. Traditional SaaS forecasting was largely based on commercial events like bookings, renewals, seat expansion, and churn. But outcome-based AI businesses behave differently. Revenue now depends on probabilistic systems, model performance, usage behavior, infrastructure cost, and constantly changing unit economics. Which means FP&A is moving from modeling commercial activity to modeling operational systems. Tommy walked through several examples of what that looks like in practice: → How “bucket” pricing structures make purchasing easier for customers, but introduce major downstream complexity across usage tracking, deferred revenue, and revenue recognition → How Fin launched its AI agent at margin-negative pricing intentionally, with the expectation that economics improve later through better routing, falling model costs, and higher resolution rates → How his team treats resolution rate itself as a financial driver, because every failed interaction still incurs inference cost while every successful resolution converts infrastructure into revenue Through each of these stories, he painted a picture of how AI-native finance teams are being rebuilt in real time. Forecasting, margin management, revenue modeling, finance tooling, and even the day-to-day workflows of FP&A teams are all changing in real time. The companies getting ahead are embedding FP&A directly into how the business operates, prices, and scales. We’re excited to continue the conversation with Fin at Beelieve London on June 4, when Aisling O'Reilly joins us to explore this same transformation from the Product point of view. Give the recap of Jeff’s chat with Tommy a read — and grab your ticket to Beelieve London — at the links below. ↓

What happens when you bring together three product operators who own pricing and ask them to challenge the industry’s thinking on AI monetization? We did exactly that at Beelieve ‘26 in San Francisco — with Jasdeep Garcha from Vercel, Zona (Xiaotian) Zhang from Clay, and Akshay Sharma of Morph Systems (formerly Miro). The discussion converged on a shift that each of the panelists is navigating deliberately: pricing is far from something you finalize after the product is built. In fact, it’s often the first product experience your customer has. Pricing is now a behavior design layer: ➡️ In Miro’s early days, they enforced a 5-seat minimum on paid plans to encourage collaboration, because a single user couldn’t experience the product’s core value alone. That constraint eventually disappeared, but it shaped how customers adopted the product early on, and it deliberately guided them to value. Pricing can also be a source of product feedback: ➡️ At Vercel, a developer’s app went viral. Instead of celebrating, he reached out with a concern: he couldn’t tell whether his spend limits were working. The team shipped a product change the same day: a clear “spend management is on” label in the dashboard, to make it easier for users to understand that Vercel’s deep spend management capabilities were enabled. And, pricing needs to evolve alongside the product and business: ➡️ Clay started with a single credit system for Data. It worked, until it didn’t. As customers began using the product for workflow automation in addition to data enrichment, one credit stopped being a clean abstraction. They introduced a second credit type for Actions to reflect that split and keep pricing aligned with where value was being created. Across all three companies, the takeaway is consistent: monetization decisions are product decisions, and product decisions are monetization decisions. -Seat minimums shape collaboration. -Credits define how usage is understood. -Spend visibility determines whether users feel safe scaling. -Model pass-through vs. bundled pricing defines where your product adds value and where you’re just reselling cost. When every interaction carries real cost, product teams can’t be insulated from commercial outcomes. More than "aligning" pricing and product, the companies that are getting ahead are building them together, as one system. This session breaks down what that looks like in practice. Read the recap and click to watch the replay at the link in comments. ↓

At Beelieve ‘26, Gorgias CFO Kunal Agarwal shared his blueprint for monetizing AI at real scale. One idea stood out early on: in AI, usage is no longer a clean proxy for value. At Gorgias, two interactions can carry identical LLM costs, but deliver completely different outcomes: resolving a support ticket vs. driving a high-value purchase. Gorgias couldn’t ignore that gap. But as they launched their AI Agent product, they also learned a critical lesson: you can’t over-engineer pricing while customers are still figuring out how to adopt AI. So, under Kunal’s leadership, Gorgias made the deliberate decision to charge only when an AI agent actually resolves an interaction. And to reduce friction in adoption, they keep pricing simple, even when they know they’re leaving money on the table (for now). Kunal shared a wealth of pragmatic insights about what it takes to operationalize AI monetization at scale, including: -What happened when he learned a single feature was quietly costing them $4 per interaction -The usage “sweet spot” Gorgias discovered that made customers far less likely to churn -Why the first 90 days determine whether AI revenue exists at all -How attribution windows impact whether customers trust your pricing -Why Gorgias reviews pricing every 90 days (and what they actually look at) If you’re working on pricing, product, or finance in an org that’s trying to monetize AI, click below to read and catch the full replay of his session. ↓ https://lnkd.in/eVaTH7wY

At Beelieve '26, we had the privilege of hosting hundreds of builders, operators, and leaders in San Francisco to share notes on how AI is rewriting the rules of modern business. In main stage keynotes from renowned thinkers Benedict Evans and Geoffrey Moore, sessions with leaders like Akshay Kothari of Notion and Kunal Agarwal of Gorgias, and panels with practitioners from Clay, Intercom, LinkedIn, OpenAI, Postman, Superhuman Mail, Vercel, and many more, one idea kept surfacing: The best teams have fundamentally shifted how they think about monetization. It’s no longer a layer, a short-term project, or a single step. It’s the system that defines how your product (and your business) grows. Across two packed days, we saw how this shows up in practice: -Product teams are embedding monetization directly into the product experience. Usage-based limits, credit systems, feature gating, and in-product upgrade paths shape how users discover and expand value from day one. -With AI, users are reaching value much faster than before — sometimes in a single session. That puts pressure on monetization to keep up: pricing, entitlements, and expansion paths need to adjust in real time based on usage, not quarterly packaging decisions. The teams that can ship and iterate on these changes quickly have a clear advantage. -Growth is more than product-led. It’s momentum-led. Teams are redesigning onboarding to get users to a meaningful “first outcome” faster, then use usage signals, limits, and prompts to drive expansion naturally from there. -And this isn’t just a product shift. Finance and operations teams are critical partners in building the infrastructure behind this: systems that can handle real-time usage tracking, flexible packaging, and constant iteration without breaking reporting, billing, or revenue recognition. We closed out Beelieve on a high note with our first-ever $100K AI Startup Pitch-Off, in partnership with Adam O'Donnell of The Founder Initiative. Congratulations to Chandrika Maheshwari and Tanay Agrawal of Quivly AI, who took home the grand prize. We look forward to following your growth journey. Thanks to the incredible lineup of speakers, customers, and partners who made Beelieve happen. Follow us here as we share more takeaways over the coming weeks. And join us at our next stop, Beelieve London, on June 4!

Having a free plan or free trial used to be a PLG growth tactic. In the world of agents, it's an imperative. Here's why: 1/ Claude isn't just evaluating your tool anymore. It's deciding on it and using it on behalf of the user. The evaluator is no longer a human skimming your landing page. It's an agent making a real-time call on whether to pull your product off the shelf and actually run it. 2/ Agents are laaaaaazzzzzyyyy. Like moths to the flame, they will choose the path of least resistance to get the job done. If your signup requires a credit card and a sales call, and your competitor's tool can be spun up in ten seconds without one, guess which one the agent picks. Signup friction isn't just a conversion tax anymore. It's a reason to get left out in the cold entirely. 3/ Once the agent has built confidence in your product, it becomes the one suggesting and evaluating the upgrade. The free tier isn't the end of monetization. It's the top of the funnel inside the agent's reasoning. Trust gets built in the free motion, and then the agent surfaces the upgrade to the human when it hits a limit that matters. 4/ Detailed packaging info is how agents evaluate you. The more clarity on your pricing and packaging page, and especially in your docs, the more likely an agent surfaces your product as the right fit. This doesn't mean publishing dollar amounts for every plan. It means making it obvious what each plan does, who it's for, what other types of customers use it, and when you'd graduate. Two experiences, one personal, one work made this real for me: I asked Claude to find two songs on my hard drive and build a mashup, end to end. It found two songs from my music folder, gave me a selection of cloud-based DAWs that didn't require a credit card, and constructed the mashup entirely on its own. (The song kinda sucked. Taste is still the human moat!) And at work, I prototyped a new customer health dashboard, and Claude picked every dev tool I needed to build and host it. Framework, hosting, deploy path. All of it. In both cases, Claude chose the tools. Not me. This is a topic I got into with Guillaume "𝑮" Cabane, former CMO of Ramp at the Chargebee Beelieve conference last week. The buyer in B2B SaaS is no longer just the human. It's the agent standing in front of the human, and the agent reaches for whatever is easiest to try, fastest to evaluate, and clearest to reason about. If there's no way for an agent to pick up your product, use it, and earn its trust without a credit card, you're invisible in the buying loop that actually matters. Excited to share more from the talk soon!

It was a pleasure!