Resources

A SaaS company added an AI summarization feature and priced it at $49 a month as an add-on. The number came from looking at what competitors charged, not from what the feature cost to run. Six months later, a subset of power users running the feature dozens of times a day were costing more in model API calls than the add-on brought in, and the finance team only found out when someone finally split the infrastructure line apart to see what was in it.

A marketplace can grow gross transaction volume 40% in a quarter and still miss its revenue target. Volume and revenue aren’t the same number, and a lot of marketplace models quietly treat them like they are. GMV is what flows through the platform, and the company only keeps a slice of it. That gap is where most marketplace forecasting goes wrong.

A controller closing the books used to spend the better part of a week matching bank feeds against invoices by hand, chasing down a $340 discrepancy that turned out to be a duplicate charge nobody caught in week one. That kind of work is where AI has made the most real difference in finance so far. Not in replacing judgment, but in cutting down the hours spent finding the thing that judgment then has to fix.

Diligence in 2026 takes longer than it used to, and investors are spending more of that time in different folders than they were two years ago. Customer contracts and usage data get more attention now than the pitch deck. A financial model that looked sufficient at Series A in 2022 gets picked apart line by line before a term sheet gets signed.

A vendor evaluation spreadsheet used to have a competitor’s name in every row. Increasingly, one row says “build internally” with an engineering estimate next to it, and that estimate is almost always too low. This is the objection SaaS companies are running into more often in 2026, and it doesn’t behave like a normal competitive threat. The buyer isn’t comparing you to another vendor with its own sales team and its own pricing page. They’re comparing you to a weekend project their own engineer is confident they can pull off with an LLM and a few days of focus.

Runway math tends to feel fixed until you actually sit down and pull it apart. A company with eleven months of cash left often assumes the only real lever is the size of the next raise. In practice, most startups have four or five smaller levers sitting in the model that nobody has touched, and pulling even two of them can turn eleven months into sixteen without changing a single thing about the fundraising environment.