Imagine sending your quarterly investor update, feeling good about it, and having someone who reads dozens of these say: “This is fine, but it reads like you’re writing to a bank, not to people who backed you.”
That’s a common reaction to AI-drafted investor updates. The second paragraph reads: “We made meaningful progress this quarter across core product development and customer acquisition initiatives.” Technically accurate — something real shipped, customers were signed. But it tells investors nothing. It could describe any company doing anything.
The problem is predictable: AI drafts the structure, you fill in some of the numbers, and you leave the transitions, the framing, the summaries. And those are exactly where the AI softens everything into corporate mush.
Try editing this investor update
A typical AI-drafted update. The structure is there but the real story isn't.
The problem isn’t that the AI got the facts wrong. It’s that it removed the facts and replaced them with the shape of facts. “Meaningful progress” has the same grammatical form as a real update, but carries no information. “Broadly in line with expectations” is something you say when you don’t want to say the number.
Investors can tell the difference. Most of them have read hundreds of these. They know what an update written to manage perception looks like versus one written to actually report. The hedged language, the vague categories, the absence of a single specific number — these are signals, and not the signal you want to send.
Why the stakes are higher than they seem
Investor updates are one of the few places where being direct builds more trust than sounding polished. An update that says “we burned $48k last month, which gives us 9 months of runway, and here’s what we’re doing about it” reads as more credible than one that says “we are monitoring our burn rate carefully.” The first one sounds like someone who knows their business. The second sounds like someone who doesn’t want to say the number.
Founders sometimes use corporate language because they’re nervous about delivering bad news. But investors fund founders partly because they believe the founder will tell them what’s actually happening. A hedge-everything update undermines that belief, even when the underlying numbers aren’t bad.
What the fix looks like
The fix is going sentence by sentence through the AI draft and replacing the summaries with the actual details.
“The first paragraph says ‘meaningful progress on customer acquisition’ — say we signed two customers, name the category they’re in, and say what each is paying.”
“Remove ‘positioning the company for continued growth’ — that sentence means nothing, cut it entirely.”
“The product paragraph says we’re ‘addressing feedback from early users’ — replace that with what we actually built and what the specific complaint was that prompted it.”
Each of those takes a few seconds to say out loud. The result is an update that still has the structure the AI gave you — the sections make sense, the order is logical — but the sentences inside it are real ones.
What AI can and can’t do here
AI is useful for the skeleton of an investor update: what sections to include, how to sequence the narrative, what to cover in each paragraph. That structure is real work and it’s genuinely helpful to not start from a blank page.
What it can’t do is fill in the specifics. It doesn’t know your actual burn rate. It doesn’t know that the two customers came from the same conference and that’s interesting because it suggests a channel worth doubling down on. It doesn’t know that “addressing feedback from early users” means one specific customer told you the onboarding was broken and it took three weeks to fix.
Those details are what the update is actually for. The AI draft is a placeholder for them, not a substitute.
The structure the AI gave you isn’t the problem. The problem is that the facts haven’t been put in yet.