Two-Week Diligence Rescue for an AI Marketing Scale-Up's Series B

The situation

A rapidly scaling AI-driven marketing startup raising a $15M Series B was in a time crunch. Investors were pushing back hard on the accuracy of projected user growth.

Growth projections are the standard pressure point at Series B, because that is where the valuation is concentrated. Investors are not usually disputing the direction; they are testing whether the company can show what the projection depends on.

The time pressure changes the problem. A funding process has momentum, and momentum has a cost — a delay to gather evidence reads as a lack of evidence, whether or not that is true.

What we did

Dug into customer data to validate growth claims via LTV and churn. Validation has to come from customer behavior rather than from the projection's internal logic. A model can be arithmetically sound and still rest on a retention assumption the data does not support.

Built a focused model showing how sensitive projections were to key assumptions. Sensitivity analysis reframes the conversation. Instead of defending a single number, it shows which assumptions the outcome actually depends on and by how much — which is the question the investor was asking.

Ran a daily cadence answering investor questions in real time. Daily rather than weekly, because in a compressed process the response time is itself evidence about how well the company knows its numbers.

What changed

The diligence summary was delivered in two weeks, and the $15M Series B closed on its original timeline.

The funding timeline held. Investors' concerns were addressed head-on rather than deflected. And better internal tracking was established to prevent a repeat.

Holding the timeline is the outcome with the most value attached. A round that slips does not merely take longer; it invites a re-examination of terms in a market that may have moved, and it signals something to every participant in the process.

What this means for companies in diligence

Investor pushback on projections is a request for the derivation, not a rejection of the number. Companies that hear it as the latter tend to defend; companies that hear it as the former produce the workings and the objection dissolves.

Sensitivity analysis is the most efficient response available. It concedes the honest point — that the projection depends on assumptions — while demonstrating exactly which ones matter, and that concession is what makes the rest credible.

The internal tracking built afterwards is the part worth keeping. A company that had to reconstruct its retention data under time pressure has learned precisely which instrumentation it was missing.