How Macabacus Monetized Its First AI Product

Case study·Aug 24, 2026

Macabacus has been selling software to finance teams since 2008, most of that time as a standard SaaS company with fixed costs and fixed prices. Its customers are predominantly financial institutions and banks. Then it built its first AI product and ran into a new problem: AI has a variable cost per use, and the company had no way to monetize it.

Rahul Gill, who leads customer experience at Macabacus, talked with us about what it took to bring usage-based billing into a business that had never needed it.

Metering has to be accurate

When you pass token costs on to the customer, the number on the invoice has to be right. That accuracy is what prevents billing disputes, and it is what finance and billing teams need for a clean ledger of costs and token burn. Macabacus evaluated building the metering itself and decided against it.

The biggest thing that was top of mind was strong capability around metering and accurately, keyword accurately, tracking token costs, because that's eventually what we're passing on to the customer. We wanted to avoid having our engineers manage a metering system. There's really no appetite for that here. That's why Schematic really stood out for us.

Predictability beats pure pay as you go

Some of the billing systems Macabacus looked at could only charge per token, with no ceiling: every token burned adds cost, indefinitely. Its customers are banks working to get AI spend under control, so Macabacus landed on credits instead. Usage is pooled and quoted on a per-seat basis, since that is still how its buyers think about value, with bundles and top-ups to get through the month.

You want to look at things like usage caps, bundles to buy more, top-ups. You need to be able to give the customer some sort of predictability around cost. That's really what they're looking for.

Agility was the deciding factor

Nobody has a consensus answer yet on how AI products should be priced, including whether per-seat billing survives a world where agents drive the usage. That makes the ability to change course more valuable than any particular pricing decision. Standard deals run on automatic plans, and custom plans spin up for the negotiations that need one.

Thousand percent, that was the key thing. Due to a special negotiation with a customer, we could change the package overnight. Having a system that actually accommodates that is huge.

Bringing in a specialist

Rahul's advice for a SaaS company approaching its first AI product: do the diligence, because usage-based billing means different things to different vendors, and the large billing platforms are still working out what it looks like out of the box. Get finance, sales, marketing, and CX in the room, because a pricing package nobody can sell is no package at all. And expect the pricing itself, not the infrastructure, to be the hard part.

You've been able to advise us from day one. It really felt like we're bringing in a specialist, which was huge, especially given a lot of this is new to our team.

Watch the full conversation: How Macabacus Monetized Its First AI Product.