Ace creates an AI proposal system that turns discovery calls into signed contracts

Three months ago we moved off PandaDoc. And off the stack of no-code tools we'd built our proposal flow on for years.

Those tools were good to us. We're an Airtable partner, we build on no-code every day, and that stack got us a long way. But we hit a ceiling, the one every off-the-shelf tool shares: it starts from a blank page. And the blank page was the problem.

What a proposal actually is

It's the moment everything you learned about a client either makes it into the deal or quietly falls out of it. For us, that "everything" is the discovery: AI interviews run across the client's whole org, mapping exactly where their work is breaking and what it's costing.

PandaDoc can't see that. No tool you buy can. They weren't in the room.

So we built our own. And it doesn't start from a blank page. It starts from the conversation.

How it works

A discovery call finishes. From there:

1. The transcript runs through our system. A proposal drafts itself: the problem, the scope, the projected gains, the phases. All of it pulled from what the client actually said, checked against our data from 311 projects.
2. A Slack notification fires to our builders. They review the scope, price the work, and finalize the document.
3. The client gets one link. They sign.
4. Signature triggers the invoice in Stripe or Xero, books it to the right codes, and hands the job to delivery.

Conversation in. Signed contract out. One system, end to end.

From 90 minutes to 4

Every proposal we sent the old way, stitched across six tools, cost us about 90 minutes of human work. Every time.

In the new system it takes 4.

That's 86 minutes of work, per proposal, that has stopped existing. Across the 69 proposals we've run through it since launch: roughly 99 hours back, about two and a half working weeks we're never spending again.

Not saved by faster software. Removed, because that work no longer needs a human at all.

This is the part people miss about AI at work. The win isn't generating a document faster. It's that the whole path from "we understand your problem" to "we're under contract solving it" collapses into one thread, and the parts that used to eat someone's afternoon just stop existing.

Three months in

Every proposal we've sent in the last three months has come out of it. 69 of them. Including the ones we sent to The Economist and Naked Wines. It's not a prototype we're piloting. It's the only way we do business now.

We didn't set out to build proposal software. We're not a proposal software company and never will be. We build operational systems that scale, and this was the one we needed, the thing no tool on the market could do for us.

The proposal used to be where good discovery went to die. Now it's what discovery produces.

First of a few I'll write on the tech we've built. More soon.